diff --git a/tools/debug-journal.md b/tools/debug-journal.md
index 78195d738..1fe1b1f7d 100644
--- a/tools/debug-journal.md
+++ b/tools/debug-journal.md
@@ -113,13 +113,13 @@ Grep for the named symbol rather than trusting a line number.
 | HA write/verify (`ha.py`, `inverter.py`) | `get_state(refresh=True)` is a no-op in a normal HA add-on install: it only re-reads when `not self.ha_key`, and `ha_key` is set in that case, so verification reads the websocket cache, never a fresh value (`ha.py:798`, since PR #2342). Separately, `write_and_poll_value()` treats `domain == "sensor"` as Predbat-owned and POSTs `/api/states/<entity>` directly (`inverter.py:2155`, `ha.py:1071-1077`) instead of calling the integration. Probed empirically across all four write-path/domain combinations (GH#4738): a control entity that resolves to a `sensor.*` domain by mistake gets a direct state overwrite that reads back as success, with zero corresponding lines in the real integration's log. `call_service_wrapper()`'s return value used to be discarded at every `inverter.py` write site, so a rejected HA service call (e.g. a Modbus "Illegal Function" error) never reached the verify logic. That is no longer true everywhere: PR #4878 made `call_service_template()` check it (`inverter.py:2817`) and log "was not accepted, it will be retried next cycle". The other eleven `call_service_wrapper()` call sites in that file still discard it, so check the specific write path before repeating the general claim. Later triage extended this row rather than replacing it. `write_and_poll_switch()` builds its service name as `domain + "/turn_..."` straight from the entity's own domain with no validation (`inverter.py:2102`), and `adjust_inverter_mode()` routes every `has_ge_eco_toggle` inverter through it - so an `inverter_mode` pointed at a `select.*` entity produces `select.turn_on`, an HA-invalid service retried 10x at 10s per attempt, and the string coercion just below (`inverter.py:2110-2111`) maps any select state to False, so the force-export direction reports success having written nothing (GH#4909, still live — though on 3-phase GEC the real problem is one stage deeper: no eco register exists at all, and the write path is skipped with `No entity_id for ECO Toggle`; see the GE Cloud row). `adjust_force_export`/`adjust_charge_window` format their times at plan precision with no snapping (`TIME_FORMAT_HMS`, `inverter.py:2608`/`:2615`), so against an integration whose select only offers 15-minute options (Solar Assistant/Deye) Predbat writes e.g. `20:55` and `write_and_poll_option` retries the impossible value 10x every cycle; nothing anywhere reads a select's `options` attribute, and `charge_time_entity_is_option` only decides *how* an entity is written, never which values are legal. The fix precedent is already in tree - AlphaESS's `snap_time_grid()` (`alphaess_const.py:215`), snap start forward and end backward with a collapse check (GH#4947). A second, independent way plan-precision times fail on the same select path: `write_and_poll_option()` only strips the seconds from an 8-char `HH:MM:SS` write when the *previous* state is a healthy 5-char `HH:MM:SS` read — the read feeding it returns `None` for an entity missing from the state cache (`ha.py:817`) or the literal `"unknown"`/`"unavailable"` when degraded (`ha.py:815`), either skips the strip and the raw `HH:MM:SS` goes to `select_option`, which HA rejects; all 10 retries re-read and fail identically, and the log reads `didn't complete got unknown` (GH#5009, probe-verified on main). That degraded-read shape can also be **permanent**: an SA slot select can report `unknown` for days — every write read-back `unknown`, zero verified writes, across an SA rebuild and both native-API and MQTT transports — so both symptoms (read path `unable to read Export window - ... returned no data` every cycle via `time_string_to_stamp("unknown")` → `None`, and the write path above) fire continuously; state reporting varies by SA backend (GH#5214). **A user edit of `config.py`'s INVERTER_DEF `charge_time_format` to anything ≠ exactly `"HH:MM:SS"` is a masking workaround, not a fix (GH#5214):** the `!= "HH:MM:SS"` branch (`inverter.py`) swaps apps.yaml entities for Predbat's own `sensor.predbat_{type}_{id}_*` dummies — reads parse Predbat's own echo and writes become direct `set_state` POSTs that never reach the inverter — so the log goes quiet while SA window control is silently severed. Tells, all observed in #5214's attachments: a `Creating dummy entity sensor.predbat_{TYPE}_{id}_charge_start_time` line in the log (`create_entity()` only logs when the entity is missing from HA state, so its presence proves the ≠ branch ran); the debug yaml's live `args` naming dummy ids while `args_from_apps_yaml` still lists the real `select.*` entities; and warnings stopping after a user restart then resuming after an addon update (which overwrites the edited `config.py`, wiping it). Also: pre-#3533 SA shipped `charge_time_format: "S"` (dummy time entities, no SA select wiring), so "SA time warnings started after updating" can mean the user crossed the #3533 template wiring rather than a format regression. The fix that serves both is reading the select's `options` attribute — nothing in the codebase does today. Six of the eight `adjust_*` writers discard the write helper's boolean and fire `call_notify()`/`mqtt_message()` regardless, and `rest_setReserve()` has the same gap (GH#4845). `call_service_template()` used to record its dedup hash *before* the call, so a silently failed `select_option` was skipped as "previously called" on every later cycle; fixed (it now records only on success and drops the record on failure), though it still deliberately returns True so callers cannot downgrade a freeze into a stop (GH#4876). From that same report, a false-alarm class worth recognising: a `didn't complete got X` where X is exactly the *previous* cycle's target, on an integration that republishes its entities slowly, is the verify window reading HA's cache - the write had already landed by the next cycle. That window is arithmetic, not a constant: up to `INVERTER_MAX_RETRY` (10) writes, each polled through one `inv_write_and_poll_sleep` budget with a doubling read interval - **~40s on GS, ~100s on GE/GEC/GEE** (`write_and_poll_sleep` is 10 for those types, 4 for GS). **Per-type since PR #5297 (merged 2026-09-29):** `INVERTER_DEF[type]["write_max_retry"]`/`["write_backoff"]` (consumed by `_write_attempts()`/`_write_backoff_result()`), and currently only `GWMQTT` deviates - 3 attempts, 10s apart - plus a per-control degraded state for those types: a write of the *same* target failing `INVERTER_WRITE_BACKOFF_FAILURES` (2) times in a row drops to one attempt per `INVERTER_WRITE_DEGRADED_INTERVAL` (300s) until a write verifies, the read-back matches or a different target arrives (full ladder restored at once); a degraded control still reads back, logs and records the failure every call - nothing is skipped silently. Every other type keeps 10 attempts and its existing sleep. A settling bounce that outlasts it is a second false-alarm shape (GH#5130): a HA `number.*` entity can step through intermediates (0 → a float → the settled int) for tens of seconds after a timed-slot write (~44s observed on solax_modbus), so a `didn't complete got <float>` whose value sits *between* the old and target values is a settling read-back, not a failed write - it self-heals next cycle when the initial read matches. The bounce reads feed only `value_matched()` (keep polling vs rewrite) and only the final read reaches the control ledger (`record_write` on match, `clear` on failure), so the worst case is a redundant same-value rewrite plus the false warning; the in-repo precedent for re-reading after settling is Solis Cloud's `verify_settle_seconds` re-read. And the discharge-stop fallback (GH#5094, still live): `adjust_charge_immediate()` issues a "stop discharge" service step before the charge/freeze step, and when `discharge_stop_service` is not configured it falls back to `charge_stop_service` with `domain="discharge"` - a documented fallback dating to #1614. On Tesla service-hook setups `charge_stop_service` maps to `self_consumption` + backup reserve 0, which on a Powerwall is **active discharge to load** (the EV charger counts as load), not a stop - so every freeze/hold cycle Predbat sends "discharge normally" and then re-commands "hold" ~2s later, and Tesla occasionally applies the first and drops the second, so the battery discharges into the car until the next 5-minute cycle. Verified from the reporter's log: the exact pair fires from the carHolding branch ("Disabling battery discharge whilst car 0 is charging" → `adjust_charge_immediate(soc_percent, freeze=True)`); the reporter had commented out `discharge_stop_service` (and `discharge_start_service`) when disabling export control, which is what exposed the fallback. Workaround: define `discharge_stop_service` in apps.yaml with commands that actually stop discharge (for a no-export setup, the same commands as `charge_freeze_service`). Caveat on the reporter's own `repeat: False` idea: enabling the service dedup means accepted calls are not re-sent (see the GH#4876 entry above - HA accepts the call but Tesla firmware can still drop it), so dropped freeze commands would never be retried. Both this and the fixed export clobber below stem from the same property of the Tesla template set: `charge_stop_service` and the freeze/discharge templates all write the same `operation_mode` select, and the templates mean opposite things on this hardware. **A sibling of the same family, over the reserve register rather than the mode selects (GH#5300, triaged 2026-09-29 against d951201e, reporter on a hand-built apps.yaml):** inside one 5-minute pass the hook charge (`adjust_charge_immediate()` issues `charge_freeze_service`/`charge_start_service` via `call_service_template`, inverter.py) and `execute.py`'s reserve reset (`if self.set_reserve_enable and resetReserve: inverter.adjust_reserve(0)`) both touch the Powerwall's reserve and the reset has **no isCharging guard** - `resetReserve` initialises True whenever `set_charge_window or set_export_window` and is cleared only by the freeze / hold-charging / car / iBoost hold branches - while `adjust_reserve()` floors at `reserve_percent` and writes through `write_and_poll_value()`, so the reset always ends up holding the register, against hook writes that are fire-and-forget. A service hook that uses the reserve register as its charge signal loses that race on every charge-window cycle. The shipped template is immune by two mechanisms, both worth checking before triaging any "Tesla reserve overwritten" report: `templates/tesla_powerwall.yaml` sets `has_reserve_soc: False` in the `inverter:` block (`fetch_inverter_data()` then force-disables `set_reserve_enable`/`set_reserve_hold`, logging `Note: Inverter does not support reserve - disabling reserve functions`), and with no `reserve:` arg wired Predbat dummies the arg (`create_missing_arg`/`create_entity`, inverter.py) so `adjust_reserve` writes land on a Predbat placeholder - the Sigenergy dummy-arg shape again. Discriminator for real-vs-dummy reserve wiring: the `Inverter N Current Reserve is X% ... and new target is Y%` and `Wrote X to reserve, successfully now Y` lines name the register, not the entity - a non-zero current read means the `reserve:` arg resolves to a real wired entity; a dummy reads 0 and at an equal target prints `already at target` instead. Dump-config tells from the same report: a `set_reserve_enable` switch showing `value: null` in a debug dump can still resolve True (expert-mode-switched item, hidden and entity not created while expert mode is off), and unknown keys in the dump's `args:` mean a hand-built config - do not assume the shipped template. | `inverter` |
 | GE Cloud (`gecloud.py`) | Gateway fields can come back null. `merge_non_null` stops nulls overwriting good values, and the publish path guards null containers — GH#4656 was a null crash in that area. `number_event` clamps an out-of-range write silently and does not report the clamped value back, while `write_and_poll_value()` keeps comparing against the original pre-clamp target - a permanent false `had_errors` and an unbounded retry (GH#4826). PR #4828 clamped inside `adjust_reserve()` only; the shared write helper still has no min/max awareness and the other `inverter.py` write sites still pass unclamped targets (GH#4831). The error-code path has since been reworked (`9560c61a`, `d6e5998c`) after GH#4896 showed `success: false` nulled `data` before the code branch could read it, so codes GivEnergy documents as never succeeding on a repeat (-3/-4/-7) were retried the full budget and `message` was never surfaced. Separately, **fixed in PR #4954 (merged 2026-09-06)** - keep the mechanism for anyone holding an older log: for `GE`/`GEC`/`GEE` the max battery rate was taken solely from the `charge_rate` entity's `max` attribute, and the `elif "battery_rate_max"` branch below it was unreachable for those types. Percentage-rated models (the 3-phase units) have no absolute charge power register, so `ge_cloud_automatic` left `charge_rate` unset and the attribute read returned the 2600 W default; both the correct rate GECloud writes into `battery_rate_max` and the user's own apps.yaml value went unread, and every rate was clamped to 2600 W. On those models it also corrupted the writes, since the percentage registers are set as `new_rate / battery_rate_max_raw * 100` against a denominator 3.8x too small. The fix falls back to `battery_rate_max` only when `charge_rate` resolves to nothing (GH#4908). GH#4198 is the same code shape on SolaX and is not covered by that fix. A model-string parser gap sits above that path (GH#5136): **the integer-rating half was fixed in PR #5293 (merged 2026-09-30, v9.3.3)** — `get_max_inverter_rate_from_model()` now parses an integer segment straight after the `3HY` token (`GIV-3HY-11` → 11 kW) *before* the decimal pass, so a later decimal segment cannot hide it; it also guards a null model, strips trailing punctuation (`GIV-HY-8.0-G3-HV.`) and warns readably on an unparseable rating ("using max charge rate instead - set inverter_limit in apps.yaml"); keep the pre-#5293 signature (decimal-only match, so an integer-rated `GIV-3HY-11` fell back to `max_charge_rate` — which the GE Cloud API can change independently, 11000 → 9984 W observed). **Still live:** `publish_info()` reads only `max_charge_rate`, never the `max_discharge_rate` in the same payload, and `battery_rate_max` is bound to the `_max_charge_rate` entity — a model whose discharge rating differs keeps the wrong rate. Distinct from GH#4908 (the absent-`charge_rate` 2600 W default). Start on any "GEC inverter rate is wrong" report by checking the model string for a decimal point; the tests now cover the integer-3HY cases too. `automatic_config()` binds `inverter_limit` to the resulting `_max_inverter_rate` sensors and `prediction.py` clips PV against `inverter_limit`, so the error reaches the plan. **Mixed fleets: the rate read/write gates test key existence, not per-slot state (GH#5359, probe-verified on main b161949d via `tools/triage_test.sh`):** `get_current_charge_rate()`/`get_current_discharge_rate()` (`inverter.py`, the `"charge_rate_percent" in self.base.args` branch) and `adjust_charge_rate()`/`adjust_discharge_rate()` (which write **both** keys, each gated only on key existence) interrogate whether the key exists **anywhere in the fleet's args**, not whether *this* inverter's slot is set — and `get_arg()` on a `None` slot applies the caller's default quietly (`userinterface.py`, float branch) — so in a fleet mixing GEC percent-controlled units (3-phase `GIV-3HY` has no `battery_charge_power` watts register; `automatic_config()` then sets `charge_rate_percent` and `build_entities()` leaves per-device `None` slots, deleting `charge_rate` when no device has the register) with watts-controlled units, a percent slot that is `None` reads quietly as 100% → `battery_rate_max_raw`, and every rate write for the power-mode sibling lands on a missing entity (`Warn: ... No entity_id for charge_rate to write ...` + `record_status(had_errors=True)`, so "Read-Only with Errors" too). The low-power rate search, which steps down relative to the read-back, then restarts from the maximum every cycle and never converges — #3311's fault extended to mixed fleets (the single-inverter shape is in the low-power symptom row). PR #4645 (open, for #3311) creates its per-slot rate dummies but does not touch these four gates — the #4908 battery-sizing read (`get_arg("charge_rate", indirect=False, index=self.id)` before falling back, inverter.py:583) is the per-entry pattern the gates lack. One API-shape trap on the site endpoint (`GE_API_SITE`, `site/{uuid}`, added for PR #4978): the published OpenAPI spec types `limits.import`/`limits.export` as a string with a null example and never shows a populated one, so the shape had to be learned from live accounts. A populated limit is `{'enabled': bool, 'power': {'watts': int, 'amps': float}}`, and **`enabled` is the thing that matters** - a site describes its connection's declared capacity in exactly the same structure as a curtailment, marked disabled, so `power.watts` alone tells you nothing about whether anything is being restricted. Site 64744 gave `{'import': {'enabled': False, 'power': {'watts': 46000, 'amps': 200}}, 'export': {'enabled': True, 'power': {'watts': 4500, 'amps': 19.6}}}` - a 200A supply whose export really is curtailed to 4.5kW - while site 46714's disabled 6kW export is its declared capacity and no restriction at all. A fleet survey confirmed the flag across accounts, so only an enabled limit is applied; an enabled zero is a real zero-export connection. A site with no limit reports a null import/export. For a fleet survey use `GET /v1/site`, the list endpoint, which returns every site's `limits` and the inverters under it (with `info.max_discharge_rate`) in one response - no second call per site. Two GEC triage tools and one clobber trap (GH#5017). The component publishes an entity for every `/settings` entry with no name filter (`async_get_inverter_settings()`/`publish_registers()`, `regname_to_ha()` is plain lowercasing), so **absence of an entity in the reporter's debug yaml is proof the API never listed the register** — "the app can control it" does not imply the settings API exposes it, because the app/GivTCP write via local control paths rather than the `/settings` list (a GIV-3HY-20 had slot 1/2 lower-SoC registers absent while 3..10 existed). Caveat: the settings list is fetched once per process start (`register_list` cache) and only the parsed flags are logged, so the debug yaml is the only observable. To capture the raw register list, `python3 apps/predbat/gecloud.py --api-key <key> --write-entity number.probe_nothing 0` makes `test_gecloud_direct()` print the full "Available entities/registers" list with raw cloud names and setting IDs and performs no write — but the harness's startup pass also runs `enable_default_options()` against the real inverter (floors reset, unused windows zeroed), so say so when asking a user to run it. Clobber trap: the missing-feature branch `set_arg("discharge_target_soc", None)` *deletes* the key, including an explicit apps.yaml entry the user wrote; `set_arg_auto(..., overwrite=False)` (component_base.py) is the in-repo pattern for letting an explicit value win — check the `has_*` detection branch on any "automatic config deleted my setting" report. The `inverter_mode` half of this class was fixed in PR #5059 (`885c637d`) — it now binds via `set_arg_auto(..., overwrite=False)` like `export_limit` — but the other missing-feature deletes are still plain `set_arg(..., None)` on main: `pause_mode`, `pause_start_time`, `pause_end_time`, `discharge_target_soc`, `charge_rate_percent`, `discharge_rate_percent`, `givtcp_rest`. A companion test-fidelity trap (GH#5056, verified on main): the component test mocks (`MockBase.set_arg` in `test_ge_cloud.py`; same shape in `test_ohme.py` and `test_alphaess_api.py`, whose mock `set_arg_auto` also writes unconditionally) store the value instead of mirroring the real `set_arg()`'s None-deletes, and their `args_from_apps_yaml` is empty so `set_arg_auto()` falls through to the plain path — so the "feature not detected" assertions (`config_args.get(...) is None`) pass identically whether the code stored None or deleted the key, and a green suite cannot distinguish fixed from unfixed. Any test for this class must populate `args_from_apps_yaml` and assert the manual value survives — #5059's Test 10 in `test_ge_cloud.py` is now the in-repo precedent. And the deeper half of GH#4909 (still live): the 3-phase GEC class exposes **no** eco/demand/mode register in `inverter/{sn}/settings` — verified by enumerating the `predbat_gecloud_*` entities in the reporter's debug dump — yet the `GEC` def sets `has_ge_eco_toggle: True` unconditionally (`config.py`) and auto-config maps `inverter_mode` via `build_entities("switch", ["enable_eco_mode"])` (gecloud.py), which emits nothing when the register is missing, so `adjust_inverter_mode()` logs `No entity_id for ECO Toggle` and returns without writing every plan cycle and freeze export silently no-ops (force-discharge paths work fine). Fix wrinkle for the maintainer: the flag is one per-inverter-type bool while `build_entities` works per-device. The freeze-export half of the same class (GH#5082, follow-up to GH#4909, code-verified on main): `inv_has_timed_pause` is probed at runtime (inverter.py flips it False when the `pause_mode` entity is absent/unavailable) but `inv_support_discharge_freeze` is read only from the static `INVERTER_DEF` entry (`config.py` sets it True for GEC), so `execute.py` never force-disables `set_export_freeze` for the eco-less/pause-less 3-phase class and a freeze-export plan can never be executed. Check both flags on any "freeze silently didn't work on GEC" report - and do not grep only for `No entity_id for ECO Toggle`: that warning comes from the eco branch, while freeze export on this class fails primarily through the pause/rate branch, so the eco warning being absent says nothing about the freeze path. Fix direction: mirror the `inv_has_timed_pause` presence-probe; the manual-freeze-export override path already degrades correctly when `set_export_freeze` is False (plan.py drops to demand). Hardware-reported on the GIV-3HY class: `charge_power_rate` limits AC (grid) charging only - `battery_power` stays at the full discharge figure while the charge register reads 1% - so the `adjust_charge_rate(0)` fallback in the freeze branch is a no-op for PV charging there; note `charge_discharge_with_rate` is False for GEC, so the first `adjust_charge_rate(0)` in each branch is skipped and the write only happens via the `has_timed_pause`-else fallback. Fix precedent for building a freeze from registers the device actually has: `inv_support_feedin_first` + Fox Cloud freeze export (PR #5038); the GIV-3HY exposes `enable_force_discharge` + `discharge_down_to_percent` + `battery_reserve_percent`, the raw material for the #4909/#5082 option-2 freeze. And since #5056 a manual `inverter_mode` apps.yaml override is silently deleted by automatic config, so "just map the eco toggle manually" is not currently viable either. The neighbouring **charge-side** gap (GH#5040: `scheduled_charge_enable` had no `enable_force_charge` candidate, so a 3-phase GEC that gates timed grid charge behind both charge switches logged real charge states while importing nothing) is **fixed in PR #5042** (`a7121135`, v9.0.2): auto-config now binds `scheduled_charge_enable` to `enable_force_charge` with `ac_charge_enable`/`enable_ac_charge` as fallbacks, and `enable_default_options()` holds `enable_ac_charge` on as a static enable when the force register exists. Keep the mechanism for pre-v9.0.2 logs, where the signature is a 3-phase GEC showing plan charge states against flat grid power with no warning anywhere. The #5042 gate itself had a second gap (**GH#5269, fixed in PR #5270, merged 2026-09-27** — keep the pre-#5270 signature): `enable_ac_charge` was written in exactly one place, `enable_default_options()` — first cycle after start-up, then once per 24h (`default_options_stamp`) — while Predbat's own `scheduled_charge_enable` write (`switch_event()`) checked nothing about the gate, so any external writer (an Axle VPP clean-up, the GE app, another integration) could clear `enable_ac_charge` and Predbat re-asserted it only on that 24h pass or a restart; "restarted and it worked" is part of the pre-fix signature, as is `enable_force_charge` on with `Charging target …%` for hours, SoC flat and no error. #5270 re-asserts the gate from `switch_event()` when a force-charge write succeeds (`ensure_ac_charge_gate`, with Axle's clean-up read as read-only for the 24h pass and a binding guard), so the "gate cleared between cache refreshes" residual window is the remaining maintainer call on that fix. So "3-phase GEC plans a charge, status says Charging, battery imports nothing" now has **two** signatures: pre-v9.0.2 = never bound (#5040), v9.0.2+ pre-#5270 = externally cleared and not re-asserted (#5269); on a post-#5270 system check whether the force switch Predbat drove is actually the `enable_force_charge` register. GE Cloud EMS plants (GH#5103, **fixed in PR #5109, merged 2026-09-24** — keep the mechanism for pre-#5109 logs): with an EMS discovered `polling_mode` is set False (`gecloud.py`), and the settings refresh used to re-read each battery inverter's `/settings` only on the first tick of the process — so the AC3 registers, including the DC-discharge slot windows, were read once and published stale for the rest of the run, while live status/meter polling kept looking normal. #5109 re-reads settings hourly (`SETTINGS_SLOW_REFRESH_SECONDS`, `gecloud.py`), refreshes the EMS and gateway devices every cycle, and runs `check_ems_inverter_slots()` on every snapshot — which now *warns* when a battery inverter's slot 1 is not the full-day window (the #3781 rule: `Inverter X slot 1 is not 00:00-23:59 ... it will override the EMS and can stop charging or discharging early`), once per episode — recorded in `ems_slot_warned` so a standing misconfiguration does not inflate the error counter on every refresh, so the non-00:00–23:59 slot 1 fault is a visible warning on current main and silent only pre-#5109 (pre-fix signature: plan/status look right but the batteries hold 0 W past a fixed wall-clock time every night). Under EMS auto-config Predbat still writes mode/schedule-enable/rates to the AC3s. Nuance: the startup read itself can still be skipped when the storage settings cache is under 10 minutes old (`settings_from_cache`), so even a restart is not guaranteed to re-snapshot the registers. **The option-validation parser used to crash-loop the whole component (GH#5217, fixed in PR #5218, merged 2026-09-26 — keep the signature for pre-#5218 logs):** `select_event()` and `publish_registers()` both `split("(")` a cloud `Value must be one of: (...)` string, so an option label containing `(` raised `ValueError: too many values to unpack (expected 2)` — the component does not die (`ComponentBase.start()` catches it), it *crash-loops*: `api_started` never set, automatic config never runs, no plan, `Error: GECloudDirect: too many values to unpack (expected 2)` every cycle. #5218 added the module-level `parse_validation_options()` (`gecloud.py`): split on the first `(` and strip only the final `)` (so inner brackets in labels survive), and `None` for a string with no `(` triggers a warning and falls back to the setting's `in:` validation-rule options instead of raising — the paren-less guard is required, not optional, because `split("(", 1)` alone fails the other direction. Which GE Cloud setting actually returns a multi-`(` validation string was never captured (the storage cache saved at the settings write would be the fixture to ask for). **`refresh_discovery()` is not device re-discovery, and GE Cloud re-discovers hourly since PR #5295 (GH#5294, merged 2026-09-29 - the symbols below were read against the pre-fix tree):** `ComponentBase.refresh_discovery()` only rebuilds the capability-catalogue *report* from data already in hand (a component opts in via `build_discovery()`) and never calls the component's API - the *similarly named* `refresh_devices()` is the thing that calls it. Pre-#5295 that device discovery (`async_get_devices()`, EV chargers, EMS/gateway selection, one-shot automatic config) ran exactly once inside `if first:` and `first` flips off permanently after the first successful run, so on any "GE Cloud didn't pick up my new inverter/charger" or "removed inverter still shows data" report a restart was the only mechanism - and a removed device kept its last good SoC/power feeding the plan (`merge_non_null()` hands back the previous reading on nulls and failed polls) and kept having registers written to. **Post-#5295:** `refresh_devices()` re-reads the device list hourly (`DEVICE_REFRESH_SECONDS`, 60*60, `gecloud.py`; a multiple of 60 because `run()` is called once a minute, and a cycle that finds a change polls and reads settings whatever else is due), `select_poll_devices()`/`apply_devices()` decide what is polled from the fresh list, and adopting a changed set drops every per-device store - `status`, `meter`, `info`, `settings`, `pending_writes`, `register_list`, the EV-charger stores, `ems_slot_warned` and `register_entity_map` - so a removed device stops publishing and being written to. The review round also: the change signature includes `battery_meters`, so a CT rewiring re-runs automatic config; a failed register-list fetch is retried on the next settings read (previously stored as `None` and never fetched again - every later read raised `TypeError`, newly reachable from rediscovery); a device read that would leave zero battery inverters is not adopted (the warning says restart if the hardware really was removed; charger-only sites stay quiet); `ge_cloud_data` returns to its pre-EMS value when the EMS leaves; the EMS slot-override warning re-arms when the EMS changes; and default options for a device discovered while read-only is set stay pending across cycles, applied once read-only ends. Caveats that survive the fix: the 5-day `last_updated` skip inside `async_get_devices()` treats a device silent over 5 days as non-functional wherever that runs - the hourly re-read included - and `merge_non_null()` still hands a *polled* device its last good SoC/power on nulls and failed polls. In-tree fix precedents (GE Cloud itself now among them): Fox re-fetches its device list every 24h (`FOX_REFRESH_STATIC`), SolaX gates plant/device-info re-reads with `data_is_due`, and `num_inverters` is re-read by the core every cycle so a re-run automatic config takes effect next plan. **The "disable" writes on the unused numbered slots leave them live (GH#5371, reporter field log from a GIV-3HY-11 Beta site, code-verified on main):** `enable_default_options()` - run at startup, once per 24h and on newly-found devices, gated only by `read_only_now()` so manual-config GEC sites get the resets too - writes every unused AC-charge and DC-discharge slot 2-10's **times** to `00:00` "to disable" (`gecloud.py`, the numbered-slot branch of the limit/time resets), and on 3-phase GIV-3HY hardware `00:00-00:00` is **live** 00:00:00-00:00:59 (minute bounds), so every "disabled" unused slot is live for one minute past midnight; with force charge on at that minute (a planned charge window spanning midnight, or the v9.1.0 stuck-on case) the unused slots fire at full rate toward their 100% upper limit (discharge mirror: full rate toward the 4% lower floor). The same pass overwrites user-set spare-slot times (the 03:00 workaround) once per pass, per the reporter's inverter-settings history log. Fix direction (maintainer's call): have the **limits** carry the disable instead (unused slots' upper→4%/reserve floor, lower→100%); the constraint is that the two generic limit branches (`*_lower_soc_percent_limit`→4%, `_upper_soc_percent_limit`/`ac_charge_upper_percent_limit`→100%) also match the numbered slots and would write straight back on the next pass, so the fix must special-case slots 2-10 inside those branches - and #5017's outcome could later drive exports through slot 3 on GIV-3HY, so the "unused" set needs a guard there. `test_ge_cloud.py`'s 00:00-reset tests stay valid under this; the fix needs numbered-slot limit assertions added. | `ge_cloud` |
 | Teslemetry / Powerwall (`teslemetry.py`) | Battery model is inferred from site `nameplate_power / battery_count`. A nameplate fallback combined with `inverter_hybrid: True` once produced a false 5 kW inverter limit and spurious morning export. Tariff writes push a whole TOU schedule every cycle, so a setting changed by hand in the Tesla app is reverted on the next cycle (GH#4600, GH#4610). Three later findings. `build_tariff()`/`_render_side` carve a *single* ON_PEAK interval priced with one scalar (`teslemetry.py:1117`/`:1036`/`:1074`), so no intra-window price gradient can exist and Tesla's own TBC is free to defer the whole export to the end of the window; the no-rate-data fallback branch also skips the boost entirely (GH#4887). The charge path sends no rate at all - `evaluate_schedule` (`teslemetry.py:539`) asserts mode `backup` plus reserve = charge target plus grid charging - so a slow charge ramp is Tesla firmware, not a Predbat write bug (GH#4892); Tesla now snaps a backup reserve of 81-99% down to 80% while Predbat still forwards any 0-100 target. The site_info limit mapping was reworked in PR #5276 (GH#5275, merged 2026-09-28): `inverter_limit` comes from `nameplate_power` (the Powerwall's own AC rating; `max_site_meter_power_ac` is the site's supply limit at the meter, not the inverter's), `export_limit` from `min_site_meter_power_ac` (kW, possibly fractional, with the ±1e9 "no limit" sentinel skipped and 0 published as a real 0 W limit), `inverter_limit_charge` at **5 kW per battery unit** capped at nameplate (Tesla reports no usable charge rating — a Powerwall 3 lists only expansion packs with every rating zero, and fleet data fits the rule), `soc_max` from the gateway's `nameplate_energy_watts` before the `battery_count` estimate, and `inverter_hybrid` now defaults from the Powerwall model (Powerwall 3 on, every other model off) with a `teslemetry_hybrid` override for a PW3 beside an existing string inverter. All of those are wired with `set_arg_auto(overwrite=False)`, so **an apps.yaml value wins** — on a pre-#5276 version `automatic_config()` overwrote a manually set `battery_rate_max` whenever `teslemetry_automatic` was on, which was the standing "my override does nothing" answer on this component. Since PR #4976 there is a Predbat-side lever for exactly that slow ramp: `teslemetry_tbc_control` — **on by default since PR #5188** (GH#5186; `DEFAULT_TBC_CONTROL = True` in `teslemetry.py`, mirrored in `COMPONENT_LIST`, with `teslemetry_tbc_control: False` as the opt-out) — switches `evaluate_schedule()` to `evaluate_schedule_tbc()`, which pushes a **signal tariff** (`build_signal_tariff()` - fixed 0/50/100p bands over the committed charge and export windows) so Tesla's own optimiser runs the charge and reaches the full rate reserve-driven charging cannot; mode goes to `autonomous` in every state, reserve becomes an actual reserve rather than a charge signal, and `_settable_reserve()` maps requests onto reserves the Powerwall honours - an 81-99% request now rounds **up** to 100 rather than being forwarded raw into the snap-to-80 band. The same PR stopped `plan.py` planning a manual freeze export on inverters that cannot do it. **Consequences of the default flip:** a Teslemetry user who never set the key is on the signal tariff with autonomous mode, so "the Tesla app shows 0p/50p/100p instead of my rates" or "my charge target isn't honoured" is the default path working, not a bug; and the 81-99% `set_reserve_min` limitation becomes default-on behaviour with it. (`MockTeslemetryAPI` still defaults `tbc_control = False` on purpose, so a green teslemetry test run exercises the real-rate path unless a test opts in.) And a separate scoping fact from the GH#5186 triage: `teslemetry_automatic` gates only `automatic_config()` — the scheduler emulator (`sync_tariff()` + `assert_device_state(evaluate_schedule(...))`) runs for every non-read-only user with no `automatic` check, so "Predbat changed my Tesla tariff/mode but I never turned on automatic" is that scoping, not a bug; `set_read_only` is the only switch that stops the writes. A related pricing bug in the same `build_tariff()` family was fixed in PR #4977: an export window ending at midnight priced the whole of tomorrow at peak. And `optimization_strategy: "economics"` *is* pushed with `tariff_content_v2` on every `set_tariff()`, unconditionally, since v8.49.0 / PR #4603 - a request to "add" it is asking for something already shipped (GH#4918). **GH#5157** (reporter's evidence, not independently probed — Powerwall 2, fw 26.26.4): the unit sat in the idle branch of `evaluate_schedule` and simply stopped acting on a standing device tuple for ~4h while `site_info` read back the *correct* configuration — drift invisible to every field the API exposes, with only a Predbat restart recovering it; the stall state is the one branch Predbat legitimately holds for hours, which is why transition-based self-heal never fires. A "Powerwall plateaued / stopped discharging until I restarted Predbat" report maps here first. **Implemented on main** (2026-09-19): `run()` re-asserts the whole device tuple — export rule, grid charging, reserve, mode, not the tariff — every `FORCED_ASSERT_SECONDS` (2h) with the write-on-change dedupe bypassed (`force=True`), the timer advancing only on a fully successful forced assert; the Info line `Teslemetry control drift-correction is transition-based ..., backed by a forced re-assert of the full device tuple every 120 minutes` names it, so on current main the stall is bounded at ~2h. **Freeze export is fully implemented in the planner and disabled for TESLA at exactly one point (GH#5225, enhancement):** `INVERTER_DEF["TESLA"]` declares `support_discharge_freeze: False` (`config.py`), read into `inv_support_discharge_freeze` (`inverter.py`), whose only consumer is the first-inverter lockstep block in `execute.py` that force-disables `set_export_freeze`/`set_export_freeze_only` (`Note: Inverter does not support discharge freeze - disabled`) — `EXPORT_MODE_FREEZE` never enters the export ladder, and PR #4976 additionally drops manual freeze-export windows to demand with a Warn. **Flipping the flag alone is not enough:** the component must also map the freeze-target discharge window to a device-side hold, or the plan's freeze assumption fails on the device — the non-TBC `evaluate_schedule()` writes the *raw* discharge target as the reserve, so a 99 target lands in the 81-99 band Tesla silently snaps to 80 (`_settable_reserve()` exists for exactly that and the non-TBC path doesn't route through it), while the default-TBC `evaluate_schedule_tbc()` discharge branch yields `pv_only` with the standing reserve but no hold — the battery still serves house load and drifts down. The device primitive exists in-tree: the TBC charge-hold tuple (`pv_only` + reserve 100 + grid charging off + autonomous) holds the battery while PV surplus exports, but a freeze *export* hold must sit at the current SoC (reserve at/above SoC through `_settable_reserve()`), not 100 — reserve 100 lets solar recharge the battery instead of exporting the surplus. Every `INVERTER_DEF` entry with `support_discharge_freeze: False` (the cloud-emulator types: SunsynkCloud, AlphaESSCloud, DeyeCloud, SolaxCloud, SolisCloud…) shares this two-part requirement — capability flag **plus** device-side freeze mapping — so enabling freeze export on any of them is a two-file change, not a one-line flag flip; each component's schedule emulator needs its own hold tuple designed against what its hardware honours (*suspected*, not verified per component). | `teslemetry` |
-| Sunsynk / DEYE (`sunsynk.py`, `deye.py`) | Freeze export is gated by the per-slot power register (`sellTime{n}Pac` on Sunsynk), confirmed on live hardware — setting the energy mode alone had no effect, and with slot power at zero the battery still charged. `read_only` was ignored by the reconcile loops until `_is_read_only()` gated `_reconcile_control()` (GH#4436). DEYE cloud telemetry keys are one hardcoded spelling per metric — `DEYE_TELEMETRY_KEYS["pv_power"] = "TotalSolarPower"` (`deye_const.py`), confirmed against a single live 3-phase hybrid — and a SUN-8K-SG05LP1-EU-AM2-P instead exposes `TotalDCInputPower` + `DCPowerPV1/2/3`, neither anywhere in the tree. Because `pv_power` is in `DEYE_TELEMETRY_REQUIRED`, the miss fires the `device/latest missing expected keys [...] - telemetry will read zero` warning every cycle and the Cloud PV sensor publishes 0.0 while a manual `pv_power` arg to a local sensor may still protect the planner. Reproduced by swapping the test fixture's single PV key (`tests/test_deye_api.py` `LIVE_DATA_LIST`). Whether the alternate key is unit-equivalent (vs summing `DCPowerPV1-3`) needs a live capture from an affected model — suspected, not verified. CONFIRMED live (Sunsynk 2211093089, 2026-09-19, via `sunsynk.py --tou-test`) that the settings object exposes the TOU register block a SECOND time under unrelated names, so several keys that read like independent settings are the same memory at a different scale: `volt1`-`volt4` are `sellTime3`-`sellTime6` as minutes-since-midnight / 50, `volt5`-`volt10` are `sellTime1Pac`-`sellTime6Pac` / 10, and `volt11`, `volt12`, `current1`-`current4` are `sellTime1Volt`-`sellTime6Volt` x 10 — a contiguous 16-value window, matched 16/16 in both the before and after reads of one write. Two consequences. It is independent proof a TOU write reached the real registers rather than a cloud-side echo, which is the cheapest hardware verification available. And it is a trap: anything that wrote `volt*`/`current*` would silently corrupt the charge programme. Predbat is safe today only because `SUNSYNK_SYSTEM_MODE_FIELDS` excludes them, so treat that exclusion as load-bearing rather than incidental. The alias does NOT extend to what the write payload must contain — `time{n}On` (capital O) is a derived echo of `time{n}on` and is correctly excluded via `SUNSYNK_DERIVED_SLOT_FIELDS`. **GH#5138**: the `control cache is X minutes old (limit 15)` forced-rewrite check is **startup-only** - it lives in `restore_state()`, called only from `run()`'s `if first:` block - so a reporter reading `limit 15` as "rechecks every 15 minutes" is wrong; it never runs again after the first cycle. The issue's core premise (idle slots never clear `sellTime{n}En`) was disproved: every slot `_owned_payload` write carries an explicit sell flag, idle slots are written with `sell: 0`, and the TOU build places a sell-0 baseline segment at each window's end (probe-verified; a `sellTime5En: 1` read-back just after an export window can be *correct*). The real gap it exposed - a write the cloud API acknowledged but the dongle never collected was trusted forever, because the owned-field change gate suppresses identical re-sends and `note_settle` only warned - is **fixed in PR #5142 (merged 2026-09-19)**: sustained settle divergence now clears `applied_payload[sn]` so the next write cycle re-applies. Keep the mechanism for pre-#5142 logs, where a stuck export slot needed a restart to clear. No debug yaml was ever attached, so the reporter's exact overrun is unconfirmed; the `has not applied Predbat's settings after N settings polls` warning is the tell. Two component-side traps since. **PR #5150** (merged 2026-09-19): `control_active` is now persisted and restored alongside `applied_payload`; **PR #5360 (merged 2026-10-05, fixes GH#5349) then split the bounds — keep the single-15-minute-bound reading for pre-#5360 (≤ v9.3.5) logs** — with `applied_payload` restoring only when its cache is ≤15 minutes old (`SUNSYNK_RESTORE_MAX_CONTROL`/`DEYE_RESTORE_MAX_CONTROL`) and `control_active` (the reconcile gate) when ≤ **8 hours** old (`SUNSYNK_RESTORE_MAX_CONTROL_ACTIVE`/`DEYE_RESTORE_MAX_CONTROL_ACTIVE`), so a restart **inside** the bound no longer strands an armed inverter in silence; on the upgrade path a cache carrying `applied_payload` with no `control_active` key infers the missing half from the payload keys — a safe lower bound, since `applied_payload[sn]` is written only by `apply_settings`/`apply_dynamic_control`, whose production callers arm `control_active` first (the CLI paths apply with `force=True` and never save). And the in-code comment on the Sunsynk write-button handler claiming Predbat "presses this on every cycle" is wrong about frequency: both press sites in `inverter.py` are change-gated (target SoC presses only when `current_soc != soc`, the export window only when `schedule_changed`), and the one unconditional post-restart re-commit is `is_hm_format`-only, which excludes these types' `HH:MM:SS` — so after a restart with the plan already matching the inverter, nothing is pressed for hours. Read the press gates in `inverter.py` before using that comment to dismiss a "component stopped writing after a restart" report. **GH#5349 (live, 2026-10-02) joins the two halves end to end:** an HA outage restarted Predbat with the control cache 35.9 minutes old, restore logged `Info: Sunsynk control cache is 35.9 minutes old (limit 15), forcing a rewrite` and dropped both halves — but nothing performs that rewrite on its own: `_reconcile_control()` gates on `control_active`, `number_event` reaches Predbat only as in-memory `update_local_schedule()` updates (no apply), and both press sites stayed change-gated, so **zero settings POSTs fired from 00:59 to 03:02** (counted from the reporter's full-API-logging log) while Predbat re-armed the hold every 5-minute cycle (reserve chasing SoC+1) and the inverter ran its 00:59 programme's idle-cap-5 slots, draining ~9 kWh (82% → 24%) into the EV until a charge-window plan change finally pressed the button at 03:02 — recovery then ran #5142's settle detection (`Warn: Sunsynk ... has not applied Predbat's settings after 4 settings polls`, fired at 03:59) and re-applied. Reading the "forcing a rewrite" line as if a rewrite follows is the trap: it is the notice that the cache was **dropped**. So a "holds lost for hours after an HA blip / battery drained while Predbat showed holds armed" report maps here on **pre-#5360 (≤ v9.3.5)** builds; since PR #5360 the ownership cache restores up to 8 hours old (`_restore_control_state()` re-sets `control_active` from the stored list), so a restart inside 8h re-arms the reconcile gate and reconcile re-applies the programme — keep the always-lost reading for pre-#5360 logs Device-visible hold legs on these types (same log): the TOU slot cap is the only one — a hold is `cap = SoC+1` in the programme (cap 65 held the battery at 64% while the car drew from grid; cap 5 drained it toward the inverter floor), and the `discharge_rate 0` entity write is symbolic on this type, so a "wrote 0 to discharge_rate but the battery drained" report reads the cap, not the entity. Also normal for this type: Sunsynk TOU charge windows cannot cross midnight (`can_span_midnight: False`) — a 23:00–00:30 window is written 23:00–23:59 and re-rolled to 00:00–00:30 at midnight (log-confirmed; not a bug). **GH#5156** (padding truncating any TOU window crossing 04:00/08:00/12:00, `build_tou_slots()` — now shared by both components in `tou_schedule.py`) is **fixed**: padding slots now carry the state of the slot they follow and sit just after it, an inert interval that cannot end a window early (`_padding_segments()` documents the issue). The pre-fix tell was "Deye stops charging at 00:00/04:00/08:00/12:00/16:00/20:00 while the Plan page still shows charging", freshly generated each cycle — not a stale default slot. **GH#5193 (cloud component, load telemetry):** on Sunsynk M3.3.8.9 the `totalPower` field (`load_power`, `SUNSYNK_TELEMETRY`) read 0–100 W through a whole overnight cycling stretch while the battery was active — including mode-2 ("Limited to Home") stretches — and read plausibly the adjacent morning with the battery idle, so the collapse correlates with the battery being active, not cleanly with the work mode; what gates it is unresolved (suspected; the reporter's manual A/B blamed the mode flip, the log's mode-2 stretches say something else does too — a controlled live test on the affected firmware is the only settle). `dailyUsed` (`load_today`) kept counting at a plausible coarse rate through the same stretch, so a balance-derived load remains feasible — and `fill_load_from_power()` is the repair that gets nothing while `totalPower` is zero. Predbat's own export control writes the whole-system `sysWorkMode` to `selling_first` for every planned export window (`derive_control_state()`, applied in the settings payload), and the component auto-wires both load args through `set_arg_auto()`'s default `overwrite=True`, so a manual apps.yaml override does NOT win — only `sunsynk_automatic: False` opts out. The in-file docstring on main records the mode/battery-export reading plus the live counter-evidence (11.1 kWh exported in a day under "Limited to Home" with `solarSell` on; a separate live test showed battery export under mode 2 with the per-slot sell flag), so the docstring's "the mode governs battery export" claim is DEYE-inferred and not settled — a control-side alternative (stay in mode 2, drive `sellTime{n}En`) stays open. The component logs every API request/response in full in `predbat.log`, so a whole Sunsynk-cloud investigation can run from the log alone; when mining logged JSON widen the field width before trusting a truncated number (`0.4` prints as `0.` under a 14-char `substr`). **Automatic config maps a battery-side current into `battery_rate_max`, and `inverter_limit` does not bound the written rate (GH#5238, code-read on main 2026-09-25/29):** `battery_rate_max()` derives watts from `chargeCurrentLimit`/`maxChargeCurrentLimit` × nominal pack voltage (`SUNSYNK_CHARGE_CURRENT_FIELDS`, first-positive-wins — a battery/BMS-side limit; GH#5238's 280 A BMS derived 14.3 kW against an 8 kW inverter), `automatic_config()` maps it to `battery_rate_max` and `ratePower` to `inverter_limit` but **never sets `inverter_limit_charge`/`inverter_limit_discharge`**, and `inverter.py` defaults both of those to `battery_rate_max_raw` when unset — so `battery_rate_max_charge = min(raw, raw) = raw`, and that is the per-slot power the plan writes. `prediction.py` clamps only *predicted* flow to `inverter_limit`; nothing feeds `inverter_limit` into the written rate. `deye.py`'s `battery_rate_max()` docstring states the same wrong assumption outright ("inverter_limit constrains the AC side separately" while deriving from `maxChargeCurrent`), and its `automatic_config()` has the same two-arg shape — so treat any "automatic config shows/writes an impossible rate" report for a cloud component through this lens first, and a fix has to map `inverter_limit_charge`/`_discharge` (to `min(ratePower, derived battery rate)`); deriving the true inverter-side current setting needs a telemetry mapping Sunsynk does not currently expose. Workaround semantics: `inverter_limit_charge`/`inverter_limit_discharge` are safe apps.yaml overrides under `sunsynk_automatic` (the component never sets them), while `battery_rate_max` is not — `set_arg_auto(overwrite=True)` clobbers an apps.yaml value every cycle. | `sunsynk_control`, `deye_control` |
+| Sunsynk / DEYE (`sunsynk.py`, `deye.py`) | Freeze export is gated by the per-slot power register (`sellTime{n}Pac` on Sunsynk), confirmed on live hardware — setting the energy mode alone had no effect, and with slot power at zero the battery still charged. `read_only` was ignored by the reconcile loops until `_is_read_only()` gated `_reconcile_control()` (GH#4436). DEYE cloud telemetry keys are one hardcoded spelling per metric — `DEYE_TELEMETRY_KEYS["pv_power"] = "TotalSolarPower"` (`deye_const.py`), confirmed against a single live 3-phase hybrid — and a SUN-8K-SG05LP1-EU-AM2-P instead exposes `TotalDCInputPower` + `DCPowerPV1/2/3`, neither anywhere in the tree. Because `pv_power` is in `DEYE_TELEMETRY_REQUIRED`, the miss fires the `device/latest missing expected keys [...] - telemetry will read zero` warning every cycle and the Cloud PV sensor publishes 0.0 while a manual `pv_power` arg to a local sensor may still protect the planner. Reproduced by swapping the test fixture's single PV key (`tests/test_deye_api.py` `LIVE_DATA_LIST`). Whether the alternate key is unit-equivalent (vs summing `DCPowerPV1-3`) needs a live capture from an affected model — suspected, not verified. CONFIRMED live (Sunsynk 2211093089, 2026-09-19, via `sunsynk.py --tou-test`) that the settings object exposes the TOU register block a SECOND time under unrelated names, so several keys that read like independent settings are the same memory at a different scale: `volt1`-`volt4` are `sellTime3`-`sellTime6` as minutes-since-midnight / 50, `volt5`-`volt10` are `sellTime1Pac`-`sellTime6Pac` / 10, and `volt11`, `volt12`, `current1`-`current4` are `sellTime1Volt`-`sellTime6Volt` x 10 — a contiguous 16-value window, matched 16/16 in both the before and after reads of one write. Two consequences. It is independent proof a TOU write reached the real registers rather than a cloud-side echo, which is the cheapest hardware verification available. And it is a trap: anything that wrote `volt*`/`current*` would silently corrupt the charge programme. Predbat is safe today only because `SUNSYNK_SYSTEM_MODE_FIELDS` excludes them, so treat that exclusion as load-bearing rather than incidental. The alias does NOT extend to what the write payload must contain — `time{n}On` (capital O) is a derived echo of `time{n}on` and is correctly excluded via `SUNSYNK_DERIVED_SLOT_FIELDS`. **GH#5138**: the `control cache is X minutes old (limit 15)` forced-rewrite check is **startup-only** - it lives in `restore_state()`, called only from `run()`'s `if first:` block - so a reporter reading `limit 15` as "rechecks every 15 minutes" is wrong; it never runs again after the first cycle. The issue's core premise (idle slots never clear `sellTime{n}En`) was disproved: every slot `_owned_payload` write carries an explicit sell flag, idle slots are written with `sell: 0`, and the TOU build places a sell-0 baseline segment at each window's end (probe-verified; a `sellTime5En: 1` read-back just after an export window can be *correct*). The real gap it exposed - a write the cloud API acknowledged but the dongle never collected was trusted forever, because the owned-field change gate suppresses identical re-sends and `note_settle` only warned - is **fixed in PR #5142 (merged 2026-09-19)**: sustained settle divergence now clears `applied_payload[sn]` so the next write cycle re-applies. Keep the mechanism for pre-#5142 logs, where a stuck export slot needed a restart to clear. No debug yaml was ever attached, so the reporter's exact overrun is unconfirmed; the `has not applied Predbat's settings after N settings polls` warning is the tell. Two component-side traps since. **PR #5150** (merged 2026-09-19): `control_active` was persisted and restored alongside `applied_payload` under one shared 15-minute bound — **split by PR #5360 (merged 2026-10-05, v9.3.6)**: `applied_payload` keeps the 15-minute change-detection bound (`SUNSYNK_RESTORE_MAX_CONTROL`/`DEYE_RESTORE_MAX_CONTROL`) while `control_active` moved to its own ownership cache with an 8-hour bound (`SUNSYNK_RESTORE_MAX_CONTROL_ACTIVE`/`DEYE_RESTORE_MAX_CONTROL_ACTIVE`), so a restart **inside** 8h no longer strands an armed inverter in silence; on the upgrade path a cache carrying `applied_payload` with no `control_active` key infers the missing half from the payload keys — a safe lower bound, since `applied_payload[sn]` is written only by `apply_settings`/`apply_dynamic_control`, whose production callers arm `control_active` first (the CLI paths apply with `force=True` and never save). And the in-code comment on the Sunsynk write-button handler claiming Predbat "presses this on every cycle" is wrong about frequency: both press sites in `inverter.py` are change-gated (target SoC presses only when `current_soc != soc`, the export window only when `schedule_changed`), and the one unconditional post-restart re-commit is `is_hm_format`-only, which excludes these types' `HH:MM:SS` — so after a restart with the plan already matching the inverter, nothing is pressed for hours. Read the press gates in `inverter.py` before using that comment to dismiss a "component stopped writing after a restart" report. **GH#5349 (live, 2026-10-02) joins the two halves end to end:** an HA outage restarted Predbat with the control cache 35.9 minutes old, restore logged `Info: Sunsynk control cache is 35.9 minutes old (limit 15), forcing a rewrite` and dropped both halves — but nothing performs that rewrite on its own: `_reconcile_control()` gates on `control_active`, `number_event` reaches Predbat only as in-memory `update_local_schedule()` updates (no apply), and both press sites stayed change-gated, so **zero settings POSTs fired from 00:59 to 03:02** (counted from the reporter's full-API-logging log) while Predbat re-armed the hold every 5-minute cycle (reserve chasing SoC+1) and the inverter ran its 00:59 programme's idle-cap-5 slots, draining ~9 kWh (82% → 24%) into the EV until a charge-window plan change finally pressed the button at 03:02 — recovery then ran #5142's settle detection (`Warn: Sunsynk ... has not applied Predbat's settings after 4 settings polls`, fired at 03:59) and re-applied. Reading the "forcing a rewrite" line as if a rewrite follows is the trap **on pre-#5360 builds** — it is the notice that the cache was **dropped**: ownership (`control_active`) expired with the payload, `_reconcile_control()` stayed gated off and zero settings POSTs fired for hours. **That hole is closed by PR #5360 (merged 2026-10-05, v9.3.6)**: ownership now survives 8 hours (`*_RESTORE_MAX_CONTROL_ACTIVE`) and `run()` calls `_reconcile_control()` on every tick that owns the inverter, so a dropped payload cache inside that bound is followed by a real rewrite; on #5360+ check the outage against the 8-hour limit first (`Info: Sunsynk control ownership cache is ... minutes old (limit 480), requiring a fresh write` = ownership expired, silent strand back until a write-button press). So a "holds lost for hours after an HA blip / battery drained while Predbat showed holds armed" report maps here on pre-#5360 versions (≤ v9.3.5). Device-visible hold legs on these types (same log): the TOU slot cap is the only one — a hold is `cap = SoC+1` in the programme (cap 65 held the battery at 64% while the car drew from grid; cap 5 drained it toward the inverter floor), and the `discharge_rate 0` entity write is symbolic on this type, so a "wrote 0 to discharge_rate but the battery drained" report reads the cap, not the entity. Also normal for this type: Sunsynk TOU charge windows cannot cross midnight (`can_span_midnight: False`) — a 23:00–00:30 window is written 23:00–23:59 and re-rolled to 00:00–00:30 at midnight (log-confirmed; not a bug). **GH#5156** (padding truncating any TOU window crossing 04:00/08:00/12:00, `build_tou_slots()` — now shared by both components in `tou_schedule.py`) is **fixed**: padding slots now carry the state of the slot they follow and sit just after it, an inert interval that cannot end a window early (`_padding_segments()` documents the issue). The pre-fix tell was "Deye stops charging at 00:00/04:00/08:00/12:00/16:00/20:00 while the Plan page still shows charging", freshly generated each cycle — not a stale default slot. **GH#5193 (cloud component, load telemetry):** on Sunsynk M3.3.8.9 the `totalPower` field (`load_power`, `SUNSYNK_TELEMETRY`) read 0–100 W through a whole overnight cycling stretch while the battery was active — including mode-2 ("Limited to Home") stretches — and read plausibly the adjacent morning with the battery idle, so the collapse correlates with the battery being active, not cleanly with the work mode; what gates it is unresolved (suspected; the reporter's manual A/B blamed the mode flip, the log's mode-2 stretches say something else does too — a controlled live test on the affected firmware is the only settle). `dailyUsed` (`load_today`) kept counting at a plausible coarse rate through the same stretch, so a balance-derived load remains feasible — and `fill_load_from_power()` is the repair that gets nothing while `totalPower` is zero. Predbat's own export control writes the whole-system `sysWorkMode` to `selling_first` for every planned export window (`derive_control_state()`, applied in the settings payload), and the component auto-wires both load args through `set_arg_auto()`'s default `overwrite=True`, so a manual apps.yaml override does NOT win — only `sunsynk_automatic: False` opts out. The in-file docstring on main records the mode/battery-export reading plus the live counter-evidence (11.1 kWh exported in a day under "Limited to Home" with `solarSell` on; a separate live test showed battery export under mode 2 with the per-slot sell flag), so the docstring's "the mode governs battery export" claim is DEYE-inferred and not settled — a control-side alternative (stay in mode 2, drive `sellTime{n}En`) stays open. The component logs every API request/response in full in `predbat.log`, so a whole Sunsynk-cloud investigation can run from the log alone; when mining logged JSON widen the field width before trusting a truncated number (`0.4` prints as `0.` under a 14-char `substr`). **Automatic config maps a battery-side current into `battery_rate_max`, and `inverter_limit` does not bound the written rate (GH#5238, code-read on main 2026-09-25/29):** `battery_rate_max()` derives watts from `chargeCurrentLimit`/`maxChargeCurrentLimit` × nominal pack voltage (`SUNSYNK_CHARGE_CURRENT_FIELDS`, first-positive-wins — a battery/BMS-side limit; GH#5238's 280 A BMS derived 14.3 kW against an 8 kW inverter), `automatic_config()` maps it to `battery_rate_max` and `ratePower` to `inverter_limit` but **never sets `inverter_limit_charge`/`inverter_limit_discharge`**, and `inverter.py` defaults both of those to `battery_rate_max_raw` when unset — so `battery_rate_max_charge = min(raw, raw) = raw`, and that is the per-slot power the plan writes. `prediction.py` clamps only *predicted* flow to `inverter_limit`; nothing feeds `inverter_limit` into the written rate. `deye.py`'s `battery_rate_max()` docstring states the same wrong assumption outright ("inverter_limit constrains the AC side separately" while deriving from `maxChargeCurrent`), and its `automatic_config()` has the same two-arg shape — so treat any "automatic config shows/writes an impossible rate" report for a cloud component through this lens first, and a fix has to map `inverter_limit_charge`/`_discharge` (to `min(ratePower, derived battery rate)`); deriving the true inverter-side current setting needs a telemetry mapping Sunsynk does not currently expose. Workaround semantics: `inverter_limit_charge`/`inverter_limit_discharge` are safe apps.yaml overrides under `sunsynk_automatic` (the component never sets them), while `battery_rate_max` is not — `set_arg_auto(overwrite=True)` clobbers an apps.yaml value every cycle. | `sunsynk_control`, `deye_control` |
 | Grid sign / arrow direction | `grid_power_invert` is owned by some integrations and not others. With two systems configured, one integration setting it `True` bleeds into the other's entities and inverts the arrows. The fix is an explicit `False` in the automatic config of both. | `sunsynk_config`, `teslemetry` |
 | Enphase (`enphase.py`) | Unofficial Enlighten endpoints. Accounts with MFA cannot log in at all. Discharge-to-grid schedules are required for export control. Writes need a double-submit CSRF token or return 403. Using the Enphase app at the same time can trip session limits. | `enphase_api` |
 | myenergi (`myenergi.py`) | `myenergi_automatic_zappi` (PR #4997) gates the Zappi half of automatic config; with it off, Zappis stay monitor-only and `control_active` is never set. `release_zappis()` has exactly one caller — `control_tick()` — reached only under `if self.control_active:` in `run()`, and `control_active` is latched at the first tick: every `enable_control()` refusal (zappi_control → automatic → automatic_zappi → enable_controls) returns *before* setting it, and the gate re-runs only when a fresh instance starts. So after a component restart with any prerequisite off, a Zappi already held in 'Stopped' stays Stopped and the `switch.*_myenergi_zappi_control` entity is not even republished (it is gated on `control_active` too) — the symptom for a future report is "car won't charge after I turned off Predbat" with a Zappi. Note `control_tick()` itself *does* release on read-only mode and on the control switch being turned off; it is the config-gate half that strands. Since PR #5150 (merged 2026-09-19) sunsynk/deye instead persist `control_active` across a restart (and re-infer it from a pre-upgrade cache), so a restart cannot strand an already-armed inverter there — the myenergi strand is the opposite direction (the gate never latching in the first place) and remains. Two adjacent lifecycle facts: published entities are **never removed** — `dashboard_item()`/`set_state()` are upsert-only (`output.py`, `ha.py` POST `/api/states`), so when a gate stops `publish_data()` publishing a switch the old entity lingers frozen at its last state until HA restarts, and there is no entity-removal path anywhere in the codebase; and `set_arg`/`set_arg_auto` writes survive a component-only restart, because `Components.restart()` stops and re-creates only the instance and never resets the shared `base.args` from apps.yaml — auto-wired values (e.g. myenergi's `car_charging_*`) persist until a full process restart. Distinguish the two restarts when checking arg lifecycle. | `myenergi` |
 | Wallbox (`wallbox.py`) | Component landed on main in PR #5410 (merged 2026-10-05, unreleased at the time of writing). Cloud API reimplemented from the Home Assistant `wallbox` integration and the `wallbox` PyPI library, over aiohttp. Findings from the first live runs (5 Oct 2026, a Pulsar Plus on firmware 6.7.43): sign in works with `User-Agent: Predbat`. **A charger run by an OCPP backend** (here Octopus, for Intelligent Octopus Go) reports `config_data.operation_mode: "ocpp"`, sits in status 209 Locked, and an API unlock (`PUT v2/charger/<id>` `{"locked": 0}`) is accepted but has no effect - so "the lock switch flips back" on such a charger is the backend, not a bug. On that locked charger a pause returned **HTTP 403** and a resume **HTTP 409**, from an account whose profile is super-admin: 403 on a control does *not* prove missing admin rights, whatever the HA integration assumes. `held_by_lock()` therefore sends neither while locked, and `held_by_ocpp()` keeps plan-led control off such a charger. `GET v3/chargers/<id>/ocpp-configuration` returns the backend address and the charge point credentials - **its reply contains a password, so never paste `--get` output unredacted**; no OCPP on/off write call is known. `locked` arrives as `1`/`0`, not a boolean, and `state_of_charge` is always null (a Type 2 connector cannot report it). Charger N is car N through `car_order`, which is numeric and append-only, so a status failure or a charger added later cannot shift it. Not yet verified on hardware when this was written: pause/resume on an unlocked, charging, non-OCPP charger; the real status ids for charging and paused; whether a 120s poll stays clear of HTTP 429. `python3 wallbox.py --username <email> --raw` is the capture tool (it prompts for the password). Release bookkeeping: `paused_by_predbat` and `schedule_pending` are stored through Storage *before* a pause is sent, a record is dropped only on a fresh status that settles it (unplugged, or connected and no longer paused), and each charger's control step is isolated in `per_charger()` so one charger's failure cannot block another - if "a charger stayed paused after control was turned off" is reported, read `release_one()` first. |
 | Gateway MQTT (`gateway.py`) | Control writes are MQTT commands the hub acknowledges. **The ack design arrived with PR #5220 and was refined by PR #5231 (both merged 2026-09-26)** — both postdate every earlier gateway log, so keep a publish-per-call reading for pre-#5220 logs: `_subscribe_acks()` subscribes to `predbat/devices/<id>/ack/+`, tolerating a broker that refuses it (a single warn via `_ack_subscribe_warned`; without the subscription or before any ack, `_send_control()` publishes exactly as before); once acks are seen, an identical command (same entity, command, payload) is published once per `_COMMAND_ACK_WINDOW` (30s), re-sent once after the unanswered window, and if that re-send is also unanswered `_acks_seen` drops so writes fall back to publish-every-call until telemetry confirms; `_process_ack()` matches any tracked id of the current value, a refusal outranks an ok (a multi-unit dispatch acks once per unit), and replay errors are handled; ids come from a clock-seeded monotonic counter (`_next_command_id()`, `PBAT<n>`), so they never restart at `PBAT1` after a restart, and the kept-id list (`_COMMAND_ACK_IDS_KEPT`) is pruned only after the new send is out, with an ack that arrives during a failing publish still counting. **PR #5172 is a separate, still-open implementation of the same feature** (`_subscribe_command_acks`, `uuid4().hex` ids, typed outcomes threaded through service dispatch/HTTP/inverter verifier) — its symbols are not on main until it rebases, so do not start a gateway-ack triage from its names. **Two gaps around serial-less and empty status, probed on main 2026-09-25 (GH#5227, enhancement — probed with temporary tests in `test_gateway.py`, reverted after):** (1) a sole slot publishing `serial=""` (shipped firmware 1.0.0 publishes every slot, including a GivEnergy slot whose serial discovery failed) is bound as the control target: `_needs_reconfigure()` treats `""` as a newly discovered inverter, `_needs_reconfigure()` treats `""` as a newly discovered inverter and auto-config binds it — note the last-resort branch that used to fall back to `candidate_aios or list(all_inverters)` was removed in `eaef6e61` (2026-10-04), which now defers auto-config entirely with a warn when no battery-capable telemetry has arrived (`_auto_configured` stays False so `_needs_reconfigure()` retries); an empty-serial slot that does report battery telemetry is still bound, so the trap survives on that shape, and the result is empty-suffix entities (`select.predbat_gateway__charge_slot1_start`) and commands addressed to `""`, which the firmware rejects (`dongle_serial required`) — a real fleet incident on 2026-09-16 (~22 min of failed commands); an empty-serial guard in `_needs_reconfigure()`/`automatic_config()` would close it, and #5227's proposed `dongle_count` deferral does *not* cover this state (`dongle_count == len(inverters)` when the empty slot is the only one). (2) `if len(status.inverters) == 0: return` (`gateway.py`) fires **before** `_inject_entities()` (EV chargers, gateway-online) and before `_last_telemetry_time`/`update_success_timestamp()` — on a hub where all GivEnergy slots are withheld (post-predbat-gateway#335, e.g. a single-inverter hub), every status message is dropped: EV data freezes and the component fails the 60-minute staleness test (`components.py`). Symptom pointer: "gateway EV sensors froze / gateway component unhealthy while the hub is up" → this early return. Removals staying bound and reappearance triggering re-config are already the behaviour (probed). **Probe trap:** the topology branch in `automatic_config()` depends on the *whole* visible set — a probe that adds a Gateway alongside the `""` slot drops the `""` entry via the battery-presence filter (`aios` requires `battery.ByteSize() > 0`) so it never reaches the last-resort branch; construct the exact visible set the scenario implies, and note the filter silently hides data-less slots from the count in several branches (the same mechanism can make a Gateway+2-AIO fleet collapse to AIO-direct control). Existing precedent for topology probes: `TestGatewayUnitControlBinding` (`test_gateway.py`). | `gateway` |
-| Octopus (`octopus.py`, `fetch.py`) | Intelligent Go tariffs are detected via `is_intelligent_go_tariff()`, and IOG-prefixed tariffs must be skipped when updating intelligent devices. Saving-session auto-join rebinding regressed when `joined_events` was empty (GH#4573). `octopus_slots_signature()` deliberately omits the time-drifting fields of active dispatch slots so a replan is not forced every cycle. `car_charging_threshold` is a strict fallback gated on `not self.car_charging_energy` (`fetch.py:228`, and `load_ml_component.py:348-361`) — it never runs as a second filter alongside a real `car_charging_energy` sensor (GH#4717). Saving-session reporting credits the full saving rate to every minute of the session on both rate tables (`load_saving_slot()`, `octopus.py:2786`); the slot dict has no baseline field to subtract (GH#2090) — a complaint that a saving session's reported total looks inflated starts here, not in a rate-fetch bug. A tariff with no `standard_unit_rates` link (IOG-TOU, GO) goes down `async_get_day_night_rates()`, which infers the off-peak window from 7 days of measurement TOU labels by plurality vote. On IOG those labels include ad-hoc dispatch slots, so a bonus slot recurring on 4+ of 7 nights became a permanent nightly cheap window and Predbat charged into it at the day rate (PR #4854 skips the inference for IOG). A "cheap slot Octopus has never heard of" report where the dispatch feeds are *empty* is this, not phantom dispatches — check the log for `Using off-peak windows [...] from measurement TOU labels`. TOU labels are UTC instants, so windows derived from them must be re-anchored for a local wall-clock tariff or they drift an hour at DST. Free-session slots have their own family of gates, separate from the saving-session ones and each found the hard way. `octopus_free_session` events used to be dropped when `code` was null, which is exactly how Octopus publishes auto-joined Weekend Happy Hours - fixed, the gate is now `start and end` and the log falls back to the event id (GH#4835). The free/saving-session args name **event** entities (`attribute="events"`/`"joined_events"`, `octopus.py`), and the BottlecapDave integration exposes Octoplus sessions as calendar + event pairs of which only the event entity carries those state attributes - the calendars have none at all, and both sides ship **disabled by default** in the integration. "The documented sensor name does not exist" plus a proposal to point at the `calendar.` twin is almost always the disabled-default trap, not a rename (GH#5370, verified against the integration's source): enable the event entity; the calendar twin cannot back these args. Matching calendar/event names differ by the `_events` suffix, so a pattern built from one domain's names mistranslates, and the deprecated pre-v17 names are scheduled for removal in January 2027. `load_free_slot()` bounded itself with `start_minutes < self.forecast_minutes` while rate minutes are indexed from `midnight_utc`, so any event more than `forecast_minutes` past midnight was dropped with no log line at all - fixed in `3bacc6e8`, the gate and the end clamp now both use `forecast_minutes + minutes_now` like `load_saving_slot()` (GH#4931). In the same function the start/end range was only updated on a successful decode while the apply block ran unconditionally, so an undecodable slot re-applied its rate over the *previous* slot's minute range (PR #4935). The join side has its own: the `joined_events` guard still ends in `saving_rate > 0` (`octopus.py:3542`), so `octopoints_per_kwh: 0` - a genuine free hour - yields no slot of any kind; probe-verified (0 gives nothing, 500 gives a saving slot, null gives nothing), and `git blame` puts that guard in `1b8c9136`, the fix for null-rate sessions (GH#3079), so it is an oversight rather than a deliberate exclusion (GH#4851). On dispatches, `rate_add_io_slots()`'s `location` check used to apply to *completed* dispatches too, and `rate_import` is rebuilt every cycle, so a retroactive AT_HOME to AWAY relabel silently un-stamped the cheap rate and `today_cost()` re-priced the whole day at the day rate (GH#4946) - **fixed in PR #4957**, which added `dispatch_billed_off_peak()` (`octopus.py:2977`): a completed dispatch is billed off-peak regardless of location, a straddling one keeps the location test, and the midday budget cap still applies. Keep the mechanism in mind when reading a log from before that merge, where the whole day reprices at the day rate. A neighbouring cap bug had no entry here at all: the daily low-rate slot budget was keyed on the *loop* minute rather than the slot start, so a window straddling noon drew a fresh 12-slot budget at 12:00 and stamped the afternoon cheap (GH#4950, **fixed in PR #4951**). The midday boundary itself is deliberate (`337db867`); only the keying was wrong. Worth knowing because "a cheap slot at an hour Octopus never offered" has at least three distinct causes in this row alone. Compare has its own rate-fetch gap: `download_octopus_rates_func()` (`octopus.py:2725`) reads only `standard-unit-rates`, and the newer IOG-SMB-FIX products are `four_rate_ev` - that endpoint answers 200 with empty results and the rates live on `day-unit-rates`/`night-unit-rates`. The main OctopusAPI component already auto-detects that product shape; compare never got the same treatment (GH#4921, probed against the live API). Lastly, `{dno_region}` in a tariff URL is substituted by `resolve_arg`'s `.format(**self.args)` (`userinterface.py:106-116`), so a missing `-` before the placeholder glues the region letter onto the product code and yields a plausible-looking 404 - check the literal URL in apps.yaml before believing a tariff has been withdrawn. Two later additions to the free/saving-session picture, one of them a correction. **eventType is carried only by the legacy `savingSessions` query path — the flexibility feed drops it (GH#4548, re-verified on main 2026-09-29 by reading `async_get_flexibility_events()`):** that function extracts only `code`/`startAt`/`endAt` from `customerFlexibilityCampaignEvents` and maps every saving event into `joinedEvents` with **no eventType**, so on the Direct path with an MPAN the joined-event loop's free-slot branch (`event_type == "WEEKEND_HAPPY_HOUR"`) can never fire and a joined Happy Hour falls through to the rewarded-saving-slot branch whenever `octopus_saving_session_rate` is set above 0. The legacy query carries eventType (the `savingSessions` GraphQL request names it and its map stores it) and is reached on the Direct path only as the flexibility path's own fallback (no MPAN, or the flexibility API returned no saving events). eventType remains the only sound discriminator between a Power Down saving session and a free Power Up/Happy Hour — the field GH#4851's `saving_rate > 0` problem needed and did not have — but *which query path served the events* decides whether you have it: "free Power Up priced as a paid saving session" on the flexibility path is first a question of provenance, and waiting for the BottlecapDave API before building on the newer feed is deliberate, not an oversight. The Direct join mutation is likewise still the legacy `joinSavingSessionsEvent`. And a genuinely counter-intuitive ordering constraint, worth reading before touching that function: Weekend Happy Hours are now skipped from `available_events` (they cannot be joined through the API - Octopus allocates them or the user books on the website), but **the skip has to sit after the reward/code/type maps are populated**, because those maps are built from the same events list and a joined Happy Hour looks its own type up there. Skip too early and the joined event has no type, takes the injected default reward, and the planner prices a free hour as an **80p/kWh saving session** (`c0fb4e9c`; the test fails if the skip is moved above the maps). Two rate-provenance reports from mid-September. **GH#5012** (car unplugged, plan still charged the house battery in the phantom cheap window): `octopus_intelligent_ignore_unplugged` correctly removed the planned dispatches, but the cheap price came from the integration's own rate sensor, not Predbat's stamping — parse the debug yaml and compare, per suspect minute, `rate_import_base` vs `rate_import_no_io` vs `io_adjusted`: a cheap price present in base and no_io with `io_adjusted` empty is the integration's rate sensor; present only after no_io is Predbat's `rate_add_io_slots()` stamping (#4516/#4950 family); `io_adjusted` populated means the minutes were flagged as adjusted — but **since PR #5304 (v9.3.3) Predbat's own `rate_add_io_slots()` flags the minutes it lowers too** (it used to write no marker at all; it mirrors the integration's `is_intelligent_adjusted`, never the fixed off-peak, time over the daily cap, or a cancelled car), so a populated `io_adjusted` no longer discriminates the integration's stamp from Predbat's own; the base-vs-`no_io` comparison still does, and `rate_replicate()` still refuses to copy `io_adjusted`-flagged minutes into future days. This attributes "integration feeds the wrong price" vs "Predbat stamps the wrong price" in one step. Nothing reverts `io_adjusted`-flagged minutes when the car is unplugged + ignore_unplugged is on — that fix would be an enhancement, blocked in practice by the integration not flagging stale adjusted entries (`io_adjusted` was empty in this dump). **`io_adjusted` itself was for a while wiped by the export/gas fetch (GH#5286, fixed in PR #5290, merged 2026-09-28):** `fetch_octopus_rates()` replaced `self.io_adjusted` on every call, so the export and gas fetches that follow the import fetch reset it to `{}` whenever `metric_octopus_export`/`metric_octopus_gas` were set — regression from #2826 dropping the old "only when adjust_key is set" guard — and with no markers left, `dynamic_load_car_strip_feed_rates()` had nothing to strip, so a car cancelled out of a dispatch still left it priced cheap and the house battery planned into it. The fix makes `self.io_adjusted` replaced only when the fetch carries an `adjust_key`; keep the mechanism for pre-#5290 logs. A companion fix (same PR) stops a compared tariff inheriting the live tariff's dispatch markers — each compared tariff now starts with none, and one left on the live import rates gets a copy. A user who instead wants the house battery limited to minutes the car actually charges has no clean knob (GH#5065): `octopus_intelligent_charging` off does **not** stop the stamping — the switch gates only car planning and vehicle prefs, planned slots are collected into `self.octopus_slots` regardless (gated only by `octopus_intelligent_ignore_unplugged`, which covers unplugged, not plugged-in-but-deferred) and are mostly consumed unconditionally by `rate_add_io_slots()` — since PR #5403 (v9.3.6) the expert `octopus_intelligent_consider_full` switch (default False) does supply car-needless future dispatch minutes (`octopus_surplus_minutes()`) that `rate_add_io_slots()` skips (octopus.py, the `minute in self.octopus_surplus` guard), but the gate deliberately cannot withhold the current half hour (`octopus_surplus_from()`: the car may already have charged in it) and the switch is off by default, so stamping remains the default path (GH#5424 symptom row above). `octopus_slot_max: 0` is not selective either — the cap is applied in `load_octopus_slots()` as well as `rate_add_io_slots()`, so it removes the car's charging slots too. The only complete workaround is unsetting `octopus_intelligent_slot`, which loses completed-slot tracking and Octopus car planning with it. **GH#5018** (tariff switched mid-day; tomorrow's export rates stayed on the old tariff and Nordpool estimates never appeared): Predbat re-reads Octopus rate events from HA every fetch cycle (`fetch_octopus_rates`), so stale export rates are the HA integration still serving the old contract — reload the integration. Two Predbat-side traps in the same report: `futurerate` only fills *missing* minutes (the `if minute not in rates` gate in fetch.py), so stale-but-present rates always win over Nordpool estimates; and `futurerate_adjust_auto` re-runs at every FutureRate init (one per fetch cycle) and persists via `set_arg`, so mid-switch it can persist `False` over the user's manual `futurerate_adjust_export: true` — log signature `FutureRate: No futurerate adjustment enabled, skipping futurerate analysis` repeating every cycle. The debug yaml does **not** contain the Octopus rate events (grep for `day_rates` comes up empty), so replays can't reproduce integration-served rate data; use the log's `Export rates: min/max/average` lines as the provenance check (identical min/max across days = static served data, and a fixed tariff has a fixed shape while a variable one doesn't). `self.mpan` is the **import** MPAN only — `async_find_tariffs()` sets it once from the first active import agreement's `meterPoint` and never overwrites it; an account with export has import and export as separate agreements with different MPANs, and the export one is retained nowhere (PR #4972 review). The saving-session/free-electricity GraphQL queries use `self.mpan` deliberately because those campaigns are import-side; any feature that needs to identify a *meter* cannot take `self.mpan` (PR #4972, merged 2026-09-19, adds `self.tariffs[direction]["mpan"]`). Symptom: two things that should describe different supply points coming out identical — that is this, not a deduplication bug. **GH#5144** (fixed in PR #5145): the REST tariff endpoints return overlapping `DIRECT_DEBIT`/`NON_DIRECT_DEBIT` rows for the same validity window and `minute_data()` writes each row over its range, so whichever row came last in the response won — and the order is not stable across periods, so the displayed rate could flip between variants from one period to the next ("always the higher rate" was an artefact of that ordering, not a property of the bug). `filter_payment_method()` (`utils.py`) now keeps one variant at the parse point on the component, day/night, annual and minute-data paths — preferring `DIRECT_DEBIT`, keeping rows with no `payment_method` (Agile) untouched, and leaving single-variant tariffs unchanged; null `payment_method` is the common case, so any future filter must keep nulls. **The daily cheap-slot cap is derived per account but enforced per car (GH#5215, semantics unresolved):** `get_octopus_slot_max()` (`octopus.py`) takes no `car_n` — it resolves the cap from the *account's* import tariff code via `has_six_hour_cap()` (matches `IOG-SMB` → 12) or a single apps.yaml integer — while both enforcement counters are locals (`slots_per_day` in `rate_add_io_slots()` and in `load_octopus_slots()`), and both callers run once per car, so each car draws a fresh budget: up to `octopus_slot_max × num_cars` cheap half-hours priced per day. Do **not** assume the docs' per-car sentence is simply wrong: the derivation is account-level but Octopus's own blog says "*Your car gets up to 6 hours of off-peak charging per day*" (quoted in GH#4830, the multi-car confusion thread), so the code is internally inconsistent and the intended semantics is genuinely open. Exposure is invisible outside IOG-SMB — uncapped tariffs default `octopus_slot_max` to 48 — so only capped multi-car installs can see the doubling; existing tests encode the per-car behaviour, and `octopus_slot_max` is read without `index=` so per-car configuration does not exist either way. **The IOG slot-confirmation strip removes the house's cheap rate too when a car's dispatch is cancelled (GH#5335, with the #5317 family - code chain on main 57ec7bf1):** with `octopus_intelligent_dynamic` on (default), `dynamic_load_car_check()` (`plan.py`) cancels every slot of a car that sits inside a started dispatch but is not seen charging after the confirmation grace (log: `car 0 is in a dispatch but not charging, cancelling its slots`), and `dynamic_load_car_strip_feed_rates()` (`octopus.py`) re-prices the house's view of those minutes back to `rate_max_base` when the cheapness came from the integration feed, **exempting the fixed IOG band** (`OCTOPUS_NIGHT_RATE_WINDOWS["iog"]` = 23:30–05:30) — so a house battery that had planned into the dispatch loses its cheap stamp as well, and PV10 re-prices the minutes at `rate_max` outright (prediction.py's `dispatch_gone` line, `minute > 30`). Version tell: the strip is v9.3.2+ code — a reporter's deliberate mid-log downgrade left the A/B in one `predbat.log`: v9.3.3 logged `Octopus Intelligent: removed the dispatch rate from 355 minutes of cars [0] which are not charging` while v9.3.1 kept the 6.57p dispatch rate (`Charging target 2%-100%`); workaround (semantics read, not live-tested): `octopus_intelligent_dynamic` off = no confirmation and no strip, ≈9.3.1 behaviour at the cost of #5229's low-load protection. Check the car side first, though: Predbat must *see* the car charging to confirm the slot, and a `plug_status` string (`eco`/`boost`) that never equals `car_charging_now_response: charging` leaves the car "not charging" through a real charge — check the sensor's own history before blaming the planner. **#5316 is the mid-dispatch variant, fixed in PR #5319 (merged 2026-10-03) — keep the pre-#5319 signature:** a car that stops part-way through a running dispatch half hour used to have the rest of that half hour re-priced at the day rate; the strip paths (`rate_add_io_slots()`, `dynamic_load_car_strip_feed_rates()`) now start from `dynamic_load_car_strip_from()` (`plan.py`) — the end of the half hour the car was last seen charging in, which is billed off-peak in full by Octopus — and the confirmation is never cleared, only outlived. Review-round refinements ride along: a sensor reading confirms a half hour only from 2 minutes into it, and the grace clock is not kept across a restart (saved state is judged on the skewed clock). The bounce symptom itself — the battery flipping Charge↔Demand while a car charges in bursts — is the supervision working as coded (GH#5390, log-verified): a non-"charging" reading counts after `DYNAMIC_LOAD_CAR_START_MINUTES` (3 min) inside the dispatch, the cheap dispatch rate is stripped from the car's still-future slots (house plan re-prices those minutes) and it resumes on the next "charging" reading — ground-truth with the hourly `Last hour: ... car X kWh` log lines against Octopus's own per-slot amounts before calling it a bug. | `octopus_*`, `saving_session*` |
+| Octopus (`octopus.py`, `fetch.py`) | Intelligent Go tariffs are detected via `is_intelligent_go_tariff()`, and IOG-prefixed tariffs must be skipped when updating intelligent devices. Saving-session auto-join rebinding regressed when `joined_events` was empty (GH#4573). `octopus_slots_signature()` deliberately omits the time-drifting fields of active dispatch slots so a replan is not forced every cycle. `car_charging_threshold` is a strict fallback gated on `not self.car_charging_energy` (`fetch.py:228`, and `load_ml_component.py:348-361`) — it never runs as a second filter alongside a real `car_charging_energy` sensor (GH#4717). Saving-session reporting credits the full saving rate to every minute of the session on both rate tables (`load_saving_slot()`, `octopus.py:2786`); the slot dict has no baseline field to subtract (GH#2090) — a complaint that a saving session's reported total looks inflated starts here, not in a rate-fetch bug. A tariff with no `standard_unit_rates` link (IOG-TOU, GO) goes down `async_get_day_night_rates()`, which infers the off-peak window from 7 days of measurement TOU labels by plurality vote. On IOG those labels include ad-hoc dispatch slots, so a bonus slot recurring on 4+ of 7 nights became a permanent nightly cheap window and Predbat charged into it at the day rate (PR #4854 skips the inference for IOG). A "cheap slot Octopus has never heard of" report where the dispatch feeds are *empty* is this, not phantom dispatches — check the log for `Using off-peak windows [...] from measurement TOU labels`. TOU labels are UTC instants, so windows derived from them must be re-anchored for a local wall-clock tariff or they drift an hour at DST. Free-session slots have their own family of gates, separate from the saving-session ones and each found the hard way. `octopus_free_session` events used to be dropped when `code` was null, which is exactly how Octopus publishes auto-joined Weekend Happy Hours - fixed, the gate is now `start and end` and the log falls back to the event id (GH#4835). The free/saving-session args name **event** entities (`attribute="events"`/`"joined_events"`, `octopus.py`), and the BottlecapDave integration exposes Octoplus sessions as calendar + event pairs of which only the event entity carries those state attributes - the calendars have none at all, and both sides ship **disabled by default** in the integration. "The documented sensor name does not exist" plus a proposal to point at the `calendar.` twin is almost always the disabled-default trap, not a rename (GH#5370, verified against the integration's source): enable the event entity; the calendar twin cannot back these args. Matching calendar/event names differ by the `_events` suffix, so a pattern built from one domain's names mistranslates, and the deprecated pre-v17 names are scheduled for removal in January 2027. `load_free_slot()` bounded itself with `start_minutes < self.forecast_minutes` while rate minutes are indexed from `midnight_utc`, so any event more than `forecast_minutes` past midnight was dropped with no log line at all - fixed in `3bacc6e8`, the gate and the end clamp now both use `forecast_minutes + minutes_now` like `load_saving_slot()` (GH#4931). In the same function the start/end range was only updated on a successful decode while the apply block ran unconditionally, so an undecodable slot re-applied its rate over the *previous* slot's minute range (PR #4935). The join side has its own: the `joined_events` guard still ends in `saving_rate > 0` (`octopus.py:3542`), so `octopoints_per_kwh: 0` - a genuine free hour - yields no slot of any kind; probe-verified (0 gives nothing, 500 gives a saving slot, null gives nothing), and `git blame` puts that guard in `1b8c9136`, the fix for null-rate sessions (GH#3079), so it is an oversight rather than a deliberate exclusion (GH#4851). On dispatches, `rate_add_io_slots()`'s `location` check used to apply to *completed* dispatches too, and `rate_import` is rebuilt every cycle, so a retroactive AT_HOME to AWAY relabel silently un-stamped the cheap rate and `today_cost()` re-priced the whole day at the day rate (GH#4946) - **fixed in PR #4957**, which added `dispatch_billed_off_peak()` (`octopus.py:2977`): a completed dispatch is billed off-peak regardless of location, a straddling one keeps the location test, and the midday budget cap still applies. Keep the mechanism in mind when reading a log from before that merge, where the whole day reprices at the day rate. A neighbouring cap bug had no entry here at all: the daily low-rate slot budget was keyed on the *loop* minute rather than the slot start, so a window straddling noon drew a fresh 12-slot budget at 12:00 and stamped the afternoon cheap (GH#4950, **fixed in PR #4951**). The midday boundary itself is deliberate (`337db867`); only the keying was wrong. Worth knowing because "a cheap slot at an hour Octopus never offered" has at least three distinct causes in this row alone. Compare has its own rate-fetch gap: `download_octopus_rates_func()` (`octopus.py:2725`) reads only `standard-unit-rates`, and the newer IOG-SMB-FIX products are `four_rate_ev` - that endpoint answers 200 with empty results and the rates live on `day-unit-rates`/`night-unit-rates`. The main OctopusAPI component already auto-detects that product shape; compare never got the same treatment (GH#4921, probed against the live API). Lastly, `{dno_region}` in a tariff URL is substituted by `resolve_arg`'s `.format(**self.args)` (`userinterface.py:106-116`), so a missing `-` before the placeholder glues the region letter onto the product code and yields a plausible-looking 404 - check the literal URL in apps.yaml before believing a tariff has been withdrawn. Two later additions to the free/saving-session picture, one of them a correction. **eventType is carried only by the legacy `savingSessions` query path — the flexibility feed drops it (GH#4548, re-verified on main 2026-09-29 by reading `async_get_flexibility_events()`):** that function extracts only `code`/`startAt`/`endAt` from `customerFlexibilityCampaignEvents` and maps every saving event into `joinedEvents` with **no eventType**, so on the Direct path with an MPAN the joined-event loop's free-slot branch (`event_type == "WEEKEND_HAPPY_HOUR"`) can never fire and a joined Happy Hour falls through to the rewarded-saving-slot branch whenever `octopus_saving_session_rate` is set above 0. The legacy query carries eventType (the `savingSessions` GraphQL request names it and its map stores it) and is reached on the Direct path only as the flexibility path's own fallback (no MPAN, or the flexibility API returned no saving events). eventType remains the only sound discriminator between a Power Down saving session and a free Power Up/Happy Hour — the field GH#4851's `saving_rate > 0` problem needed and did not have — but *which query path served the events* decides whether you have it: "free Power Up priced as a paid saving session" on the flexibility path is first a question of provenance, and waiting for the BottlecapDave API before building on the newer feed is deliberate, not an oversight. The Direct join mutation is likewise still the legacy `joinSavingSessionsEvent`. And a genuinely counter-intuitive ordering constraint, worth reading before touching that function: Weekend Happy Hours are now skipped from `available_events` (they cannot be joined through the API - Octopus allocates them or the user books on the website), but **the skip has to sit after the reward/code/type maps are populated**, because those maps are built from the same events list and a joined Happy Hour looks its own type up there. Skip too early and the joined event has no type, takes the injected default reward, and the planner prices a free hour as an **80p/kWh saving session** (`c0fb4e9c`; the test fails if the skip is moved above the maps). Two rate-provenance reports from mid-September. **GH#5012** (car unplugged, plan still charged the house battery in the phantom cheap window): `octopus_intelligent_ignore_unplugged` correctly removed the planned dispatches, but the cheap price came from the integration's own rate sensor, not Predbat's stamping — parse the debug yaml and compare, per suspect minute, `rate_import_base` vs `rate_import_no_io` vs `io_adjusted`: a cheap price present in base and no_io with `io_adjusted` empty is the integration's rate sensor; present only after no_io is Predbat's `rate_add_io_slots()` stamping (#4516/#4950 family); `io_adjusted` populated means the minutes were flagged as adjusted — but **since PR #5304 (v9.3.3) Predbat's own `rate_add_io_slots()` flags the minutes it lowers too** (it used to write no marker at all; it mirrors the integration's `is_intelligent_adjusted`, never the fixed off-peak, time over the daily cap, or a cancelled car — and from PR #5392 (v9.3.6) only lowerings beyond a penny (`IO_RATE_TOLERANCE`): sub-penny dp2-rounding differences no longer mark or re-price the fixed off-peak minutes), so a populated `io_adjusted` no longer discriminates the integration's stamp from Predbat's own; the base-vs-`no_io` comparison still does, and `rate_replicate()` still refuses to copy `io_adjusted`-flagged minutes into future days. This attributes "integration feeds the wrong price" vs "Predbat stamps the wrong price" in one step. Nothing reverts `io_adjusted`-flagged minutes when the car is unplugged + ignore_unplugged is on — that fix would be an enhancement, blocked in practice by the integration not flagging stale adjusted entries (`io_adjusted` was empty in this dump). **`io_adjusted` itself was for a while wiped by the export/gas fetch (GH#5286, fixed in PR #5290, merged 2026-09-28):** `fetch_octopus_rates()` replaced `self.io_adjusted` on every call, so the export and gas fetches that follow the import fetch reset it to `{}` whenever `metric_octopus_export`/`metric_octopus_gas` were set — regression from #2826 dropping the old "only when adjust_key is set" guard — and with no markers left, `dynamic_load_car_strip_feed_rates()` had nothing to strip, so a car cancelled out of a dispatch still left it priced cheap and the house battery planned into it. The fix makes `self.io_adjusted` replaced only when the fetch carries an `adjust_key`; keep the mechanism for pre-#5290 logs. A companion fix (same PR) stops a compared tariff inheriting the live tariff's dispatch markers — each compared tariff now starts with none, and one left on the live import rates gets a copy. A user who instead wants the house battery limited to minutes the car actually charges has no clean knob (GH#5065): `octopus_intelligent_charging` off does **not** stop the stamping — the switch gates only car planning and vehicle prefs, planned slots are collected into `self.octopus_slots` regardless (gated only by `octopus_intelligent_ignore_unplugged`, which covers unplugged, not plugged-in-but-deferred) and are consumed unconditionally by `rate_add_io_slots()`. `octopus_slot_max: 0` is not selective either — the cap is applied in `load_octopus_slots()` as well as `rate_add_io_slots()`, so it removes the car's charging slots too. The only complete workaround is unsetting `octopus_intelligent_slot`, which loses completed-slot tracking and Octopus car planning with it. **GH#5018** (tariff switched mid-day; tomorrow's export rates stayed on the old tariff and Nordpool estimates never appeared): Predbat re-reads Octopus rate events from HA every fetch cycle (`fetch_octopus_rates`), so stale export rates are the HA integration still serving the old contract — reload the integration. Two Predbat-side traps in the same report: `futurerate` only fills *missing* minutes (the `if minute not in rates` gate in fetch.py), so stale-but-present rates always win over Nordpool estimates; and `futurerate_adjust_auto` re-runs at every FutureRate init (one per fetch cycle) and persists via `set_arg`, so mid-switch it can persist `False` over the user's manual `futurerate_adjust_export: true` — log signature `FutureRate: No futurerate adjustment enabled, skipping futurerate analysis` repeating every cycle. The debug yaml does **not** contain the Octopus rate events (grep for `day_rates` comes up empty), so replays can't reproduce integration-served rate data; use the log's `Export rates: min/max/average` lines as the provenance check (identical min/max across days = static served data, and a fixed tariff has a fixed shape while a variable one doesn't). `self.mpan` is the **import** MPAN only — `async_find_tariffs()` sets it once from the first active import agreement's `meterPoint` and never overwrites it; an account with export has import and export as separate agreements with different MPANs, and the export one is retained nowhere (PR #4972 review). The saving-session/free-electricity GraphQL queries use `self.mpan` deliberately because those campaigns are import-side; any feature that needs to identify a *meter* cannot take `self.mpan` (PR #4972, merged 2026-09-19, adds `self.tariffs[direction]["mpan"]`). Symptom: two things that should describe different supply points coming out identical — that is this, not a deduplication bug. **GH#5144** (fixed in PR #5145): the REST tariff endpoints return overlapping `DIRECT_DEBIT`/`NON_DIRECT_DEBIT` rows for the same validity window and `minute_data()` writes each row over its range, so whichever row came last in the response won — and the order is not stable across periods, so the displayed rate could flip between variants from one period to the next ("always the higher rate" was an artefact of that ordering, not a property of the bug). `filter_payment_method()` (`utils.py`) now keeps one variant at the parse point on the component, day/night, annual and minute-data paths — preferring `DIRECT_DEBIT`, keeping rows with no `payment_method` (Agile) untouched, and leaving single-variant tariffs unchanged; null `payment_method` is the common case, so any future filter must keep nulls. **The daily cheap-slot cap is derived per account but enforced per car (GH#5215, semantics unresolved):** `get_octopus_slot_max()` (`octopus.py`) takes no `car_n` — it resolves the cap from the *account's* import tariff code via `has_six_hour_cap()` (matches `IOG-SMB` → 12) or a single apps.yaml integer — while both enforcement counters are locals (`slots_per_day` in `rate_add_io_slots()` and in `load_octopus_slots()`), and both callers run once per car, so each car draws a fresh budget: up to `octopus_slot_max × num_cars` cheap half-hours priced per day. Do **not** assume the docs' per-car sentence is simply wrong: the derivation is account-level but Octopus's own blog says "*Your car gets up to 6 hours of off-peak charging per day*" (quoted in GH#4830, the multi-car confusion thread), so the code is internally inconsistent and the intended semantics is genuinely open. Exposure is invisible outside IOG-SMB — uncapped tariffs default `octopus_slot_max` to 48 — so only capped multi-car installs can see the doubling; existing tests encode the per-car behaviour, and `octopus_slot_max` is read without `index=` so per-car configuration does not exist either way. **The IOG slot-confirmation strip removes the house's cheap rate too when a car's dispatch is cancelled (GH#5335, with the #5317 family - code chain on main 57ec7bf1):** with `octopus_intelligent_dynamic` on (default), `dynamic_load_car_check()` (`plan.py`) cancels every slot of a car that sits inside a started dispatch but is not seen charging after the confirmation grace (log: `car 0 is in a dispatch but not charging, cancelling its slots`), and `dynamic_load_car_strip_feed_rates()` (`octopus.py`) re-prices the house's view of those minutes back to `rate_max_base` when the cheapness came from the integration feed, **exempting the fixed IOG band** (`OCTOPUS_NIGHT_RATE_WINDOWS["iog"]` = 23:30–05:30) — so a house battery that had planned into the dispatch loses its cheap stamp as well, and PV10 re-prices the minutes at `rate_max` outright (prediction.py's `dispatch_gone` line, `minute > 30`). Version tell: the strip is v9.3.2+ code — a reporter's deliberate mid-log downgrade left the A/B in one `predbat.log`: v9.3.3 logged `Octopus Intelligent: removed the dispatch rate from 355 minutes of cars [0] which are not charging` while v9.3.1 kept the 6.57p dispatch rate (`Charging target 2%-100%`); workaround (semantics read, not live-tested): `octopus_intelligent_dynamic` off = no confirmation and no strip, ≈9.3.1 behaviour at the cost of #5229's low-load protection. Check the car side first, though: Predbat must *see* the car charging to confirm the slot, and a `plug_status` string (`eco`/`boost`) that never equals `car_charging_now_response: charging` leaves the car "not charging" through a real charge — check the sensor's own history before blaming the planner. **#5316 is the mid-dispatch variant, fixed in PR #5319 (merged 2026-10-03) — keep the pre-#5319 signature:** a car that stops part-way through a running dispatch half hour used to have the rest of that half hour re-priced at the day rate; the strip paths (`rate_add_io_slots()`, `dynamic_load_car_strip_feed_rates()`) now start from `dynamic_load_car_strip_from()` (`plan.py`) — the end of the half hour the car was last seen charging in, which is billed off-peak in full by Octopus — and the confirmation is never cleared, only outlived. Review-round refinements ride along: a sensor reading confirms a half hour only from 2 minutes into it, and the grace clock is not kept across a restart (saved state is judged on the skewed clock). The bounce symptom itself — the battery flipping Charge↔Demand while a car charges in bursts — is the supervision working as coded (GH#5390, log-verified): a non-"charging" reading counts after `DYNAMIC_LOAD_CAR_START_MINUTES` (3 min) inside the dispatch, the cheap dispatch rate is stripped from the car's still-future slots (house plan re-prices those minutes) and it resumes on the next "charging" reading — ground-truth with the hourly `Last hour: ... car X kWh` log lines against Octopus's own per-slot amounts before calling it a bug. **Planned dispatches on the Octopus entity are empty by design for an Ohme charger (GH#5413, BCD-source-verified 2026-10-07):** BottlecapDave's `get_intelligent_features()` special-cases OHME to `planned_dispatches_supported = False` (OHME is not in its `FULLY_SUPPORTED_INTELLIGENT_PROVIDERS`), and the integration's dispatches coordinator then clears planned dispatches outright (log: `Clearing planned dispatches due to not being supported for provider`), keeping completed and started - Octopus's own API has carried Ohme Home Pro sessions as smart-charge planned dispatches only since ~29 Sep 2026 (documented with an API probe in BCD feature request #1893, still open). So pointing a car's `octopus_intelligent_slot` at the BCD `*_intelligent_dispatching` entity on an Ohme-scheduled setup buys completed-slot tracking but zero forward car-slot visibility, and an empty `planned_dispatches` attribute is the gate's doing, not evidence Octopus holds no dispatches for Ohme. Predbat-side (fetch.py): completed merge into `octopus_slots` unconditionally, planned is gated on plugged-in/ignore-unplugged, and started is read only for the #4516 Stage 1 diagnostic - so a real fix would merge completed from Octopus with planned/started from Ohme, the approach flagged in the #5413 thread. | `octopus_*`, `saving_session*` |
 | Kraken / EDF (`kraken.py`) | EDF Kraken answers a day/night-structured tariff with **HTTP 400** on `standard-unit-rates/` (`{"detail": "This tariff has day and night rates, not standard."}`) while `day-unit-rates`/`night-unit-rates` return 200 and standing charges 200 — where Octopus four-rate products (GH#4921) answer the same-shaped URL **200 with empty results**. So the "empty results, then data empty" signature belongs to `octopus.py` and a hard 400 to `kraken.py`; do not carry the expectation across (GH#5166, probed against the live API). A nonexistent product still 404s, so 404 and 400 are both live "REST can't serve this tariff" signals on EDF — probe the exact URL from the reporter's log before assuming either. **Fixed in PR #5167** (`async_fetch_rates()`): the GraphQL `applicableRates` fallback now gates on `KRAKEN_REST_RATES_UNAVAILABLE_STATUSES` = (400, 404, 410), per direction (404 covers both a private product and one retired from the REST API; the authenticated retry stays 404/410 because auth cannot change a 400), and only counts a failure when the fallback comes back empty without counting one — `async_graphql_query()` counts its own failures, so a caller that also counts scores 2 per down cycle; snapshot the counter around the call. The standing-charge fallback stays deliberately 404/410-only (`kraken.py` — "400 is a rates-endpoint-only answer"). Two traps from the fix's review: `_fetch_rates_rest()` returns `(None, None)` on a network error, so `err is None` does **not** mean success — success is `(results, None)`; the merged code stashes the public attempt's status and restores it when the authenticated retry hits a network error (pre-fix symptom: a private-product tariff's rates vanish for one cycle with no fallback log line, while an identical 404 one cycle later recovers). The day/night endpoints' rows carry the same `payment_method` variant overlap as GH#5144 — suspected, no probe; check `minute_data`'s handling before consuming them directly. Also benign: `Warn: Kraken: Auth not available for find-tariffs` repeating on early restarts before the first `Tariff discovered` line is setup-phase, not a second bug. | `kraken` |
 | Axle (`axle.py`) | Export sessions have to boost the import rate as well as the export rate - introduced deliberately by PR #4520 (first release v8.48.2, mirrored on Octopus saving sessions), documented in docs/energy-rates.md, and upheld by the maintainer on #5060. Two reports of the same design now exist (#5060 closed as dup-of-#4277 with the by-design answer given; #5175 triaged as enhancement, no regression - `tools/triage_test.sh axle` asserts the dual boost), and **no config knob disables just the import side** (`axle_pence_per_kwh` scales both directions; config.py has only `axle_api_key`/`axle_pence_per_kwh`/`axle_automatic`/`axle_control`). The "textual plans / colour coding skewed" half of such reports belongs to the #5050 threshold row (boosts before `rate_scan()`), not here. State is published unconditionally from `run()` so a fetch failure does not freeze the sensor at a stale value. `load_axle_slot()` has no lower time bound - it checks only `start_minutes < forecast_minutes + minutes_now`, not the `start_minutes >= 0` guard `load_free_slot()` has (octopus.py) - so a closed Axle event writes dead keys at negative minutes through `rate_dict.get(minute, 0)` (GH#5036, replay-verified against the reporter's dump: the +100 boosts sit only at the two closed events' actual UTC start/end times). The dead keys are inert in the plan - `rate_minmax()` and the window scans start at `minutes_now` - so "a dead event projected forward as free export" is not what the code does. The comment at octopus.py:2859 saying `load_saving_slot()` and `load_axle_slot()` "both bound themselves that way" is half-wrong: axle bounds only its end, and `load_saving_slot()` also lacks the start guard but is harmless there because its inner loop writes only `if minute in rate_dict`. Two September reports extend this row. **GH#5060** (closed as a duplicate of #4277, the same defect reported a year earlier and never root-caused — #4277 covers Octopus saving sessions too): event boosts are applied *before* user rate overrides on both sides — `load_axle_slot(..., export=True)` runs ahead of `basic_rates(rates_export_override, ...)` (import mirrors it: `load_saving_slot`/`load_axle_slot` then the import override) — so a full-horizon `rates_export_override` rewrites the event minutes and re-marks them `user`. That reporter had no `rates_export` key in apps.yaml at all, so the override was their only export source, which is why the import-side boost survived while the export-side +100 did not. Workaround (maintainer's suggestion on #4277, checked viable against the code): move the fixed schedule from `rates_export_override` into `rates_export` — it becomes the base tariff via `basic_rates` before the boosts run, so the event stacks on top and nothing downstream rewrites it (assuming manual export rates are empty, since `apply_manual_rates` also runs after the boosts). Forensic technique: in a debug yaml, the `rate_*_replicated` dict discriminates the two failure modes in one step — mark `user` through the event window ⇒ the boost ran and a later override clobbered it; the `saving` marks missing while the boost is missing ⇒ the boost call never ran (or, per GH#5036 above, wrote dead keys). Compare against `rate_*_base` for the tariff's own price; same idea as the GH#5012 base/no_io/adjusted comparison in the Octopus row. **GH#5050** (shared with the Octopus row): the same boosts run before `rate_scan()`, so the automatic low-rate threshold classifies the whole day low — see the symptom-table row on `set_rate_thresholds()`. | `axle` |
 | History fetch / memory (`ha.py`) | History is fetched in `HISTORY_CHUNK_DAYS`-sized chunks with boundary dedup — records landing exactly on a chunk start inside a data gap corrupted smoothing before that was fixed. The largest memory peak in a run is ML load-predictor training (`load_predictor.py`), not the plan. The same chunking is the REST-call amplifier that can trip the fatal API-error counter at startup — see the "Too many API errors" symptom row. | `history_chunking` |
@@ -133,7 +133,7 @@ Grep for the named symbol rather than trusting a line number.
 | Savings & metrics (`output.py`, `predbat.py`) | `savings_total_predbat` accumulates the **unadjusted** `saving` (`predbat.py:1183`, fed from `self.savings_today_predbat = saving` in `output.py`) while `savings_yesterday_predbat` publishes `saving_adjusted` (`output.py:3380`) — the two sensors answer different questions and can disagree in sign on the same day (confirmed from a reporter's dump: `saving_real: +86.14p` vs `saving_adjusted: -25.66p`, GH#3894). Both are still current on `main`. Separately, the battery-value adjustment bills the SoC **level** at day-end against the counterfactual baseline rather than the **change** in SoC over the day — a day that force-exports through midnight can be energy-neutral yet still get charged the full baseline-vs-actual SoC gap as if it were lost value. | none |
 | GivTCP REST (`inverter.py`) | **Structurally stale as written, and left here for the mechanism only:** PR #4864 moved REST handling out into `givtcp.py`/`givtcp_rest.py`, so `update_status()` (now `inverter.py:1333`) no longer reads `Power.Power` at `inverter.py:1435` at all. What survives is the same trap one layer up - the component's auto-config claims the power keys unless `givtcp_rest_power_ignore` is set, and it now logs an Info line when you opt out (`givtcp.py:766-767`). PR #4959 also lets an apps.yaml-named energy sensor win over auto-configuration for the history-read keys. Historically, with `givtcp_rest` configured `update_status()` read PV/Grid/Load power straight out of the REST `Power.Power` block; the apps.yaml entity lists - including any `0` placeholders put there deliberately to zero a duplicate reading - are only consulted on the non-REST `else` branch, and `execute.py` then sums every inverter's REST readings. On a hybrid + AIO pair that presents as the AIO's hybrid-fed PV port counted as solar at night, and both units' shared-CT grid readings summed (~7.1 kW shown for ~3.6 kW of real export). The per-inverter `givtcp_rest_power_ignore: true` restores the apps.yaml lists and is already documented in `docs/apps-yaml.md` - the gap is that nothing warns when a `0` placeholder is silently bypassed (GH#4883). Two v9.0.x notes since the refactor. **GH#4993 (still live on main, b8996659):** the component's required `rest_urls` arg is resolved straight from `givtcp_rest` (`components.py`) and the stock `config/apps.yaml` ships `givtcp_rest:` uncommented with example URLs, while the registry's `"inverter": True` flag is documentation-only and nothing reads it - so every install using the template, GivEnergy or not, starts a GivTCP REST component that fails discovery (`num_entities: 0`), hammers two dead URLs and shows "GivTCP REST: Error / Never updated" in the Components panel. Discovered inverters stay at zero so `automatic_config()` skips and no Predbat args get hijacked - the damage is log spam and panel error, not wrong control. Workaround: comment out the whole `givtcp_rest:` block, and any uncommented `givtcp_automatic:` line with it, because leaving that set while `givtcp_rest` is absent trips the `Warn: Skipping GivTCP REST interface, missing required configuration: givtcp_rest` gate. **GH#4984 (fixed by the refactor, keep for older logs):** `Warn: Inverter N REST failed to setDischargeRate to <X> got <X>` where *got equals the requested rate* is a zero verification tolerance, not a failed write - pre-#4864 the read-back tolerance was sized from `battery_rate_max_discharge`/`charge`, which an `inverter_limit_discharge`/`inverter_limit_charge` API override of 0 (users reach for it as "manual rate 0"; it is in `CONFIG_API_OVERRIDE`) drives to 0 against a strict `<`, so an exact read-back fails, burns the 5 `INVERTER_MAX_RETRY_REST` posts (~2s apart, the 4-5 runs users see in the GivTCP log) and flags the whole run "Demand with Errors". Only fires when something else set the real rate away from the target first, otherwise `adjust_discharge_rate`'s change-gate skips the write. The refactor re-sized the tolerance from `write_tolerance_watts()` - the max rate GivTCP itself reports (`Invertor_Max_Bat_Rate`), falling back to 2600W - and the entity write path uses `<=`, so an exact match passes; the charge-rate variant #2882 (still open) does not match this trap's signature — it is a ~1 kW read-back miss (3600 written, 2560 got), not got==requested — and maps instead to the GH#5386 family appended below. **GH#5101 (still live):** `initialize()` builds one `InverterRestState(id=n, rest_api=url)` per `givtcp_rest` entry with no URL-shape validation (`givtcp.py`), so a YAML mapping indented into the list during an apps.yaml edit becomes `rest_api=<dict>` and `read_data()`'s first line (`url = inverter.rest_api + "/" + api`, `givtcp_rest.py`) raises `TypeError: unsupported operand type(s) for +: 'dict' and 'str'`, caught by `ComponentBase.start()` and logged with no index or URL. Signature: the `discovered N inverter(s) from M configured REST endpoint(s)` line never appears anywhere in the log (discovery settles only after the poll loop completes), and the malformed entry itself produces no `Errno 22` retries because it dies before `get_data()` — valid-but-dead string entries ahead of it produce the full ladder, so Errno-22 count = 5 attempts × completed poll cycles. The component never reports started; the main loop keeps working off the user's own apps.yaml entities. The whole class is new with the v9 component — v8.55 read `givtcp_rest` with `index=self.id`, so stray extra entries were inert on single-inverter installs. Also from GH#5099: nothing in the repo sets `battery_calibration` automatically — the component publishes the sensor from GivTCP's own detection but expects an apps.yaml mapping — and Predbat reads the arg's state as calibrating only when it is literally `on`/`On`/`true`/`True` (`inverter.py`), while GivTCP's detection is `Battery_Calibration != "Off"` (`givtcp_rest.py`); anyone wiring a multi-valued select into `battery_calibration` must reconcile the two. and because `refresh_config()` re-samples the arg every cycle (the objects persist since PR #5126, but the per-cycle re-read survives — see the args trap in "Traps when investigating"), `in_calibration` is re-sampled per cycle, not init-only. Another manual-vs-auto split (GH#5132): the `discharge_target_soc` write is the force-export hardware backstop - always `int(self.reserve_percent)`, never the plan's per-window export limit (that floor is software-only in `execute.py`) - and the #4517 unsupported-model guard (`DISCHARGE_TARGET_UNSUPPORTED_MODELS`/`_TYPES`, givtcp.py) runs only in the **auto-config** path, where it stops Predbat publishing the entity at all. A user on the **manual** entity-list path gets the entity from the GivTCP add-on instead, so the guard never engages: on an AIO whose add-on write handler for `number.givtcp_*_discharge_target_soc_1` is a silent no-op (diagnosed live - manual `number.set_value` did nothing and produced zero GivTCP log lines), Predbat warned `didn't complete got 4.0` every cycle for 14h. The write is skipped when the arg is absent by design, so **removing `discharge_target_soc` from apps.yaml is the first workaround on any version and either path**. Setup tell: manual-path apps.yaml lists `discharge_target_soc: - number.givtcp_*`; auto-config users don't. (Version trap from the same issue: the issue body said v8.54.3, the attached log's own version lines said v9.0.3 - the log wins.) **GH#5178 (enhancement, live on main):** whenever `givtcp_rest` is configured the component publishes its **full** entity surface unconditionally — ~44 entities per inverter — regardless of whether the GivTCP HA integration already provides equivalents; deliberate, so REST-only setups have entities at all. A "Predbat created N duplicate GivTCP entities" report maps here and is a feature request: the count is checkable as (GIVTCP_SENSORS + GIVTCP_CONTROLS − per-fleet withheld) × inverter count (44×3 = 132 in #5178, matched), the user-side workaround is an HA recorder exclusion glob `*.predbat_givtcp_*` plus hiding (safe — the four history keys keep the user's own sensors, USER_WINS), and `docs/components.md` has no GivTCP REST section yet. Related live gap (code-read, #5178's reporter's debug yaml): `automatic_config()` claims `discharge_target_soc` gated on rest_v3 alone (`givtcp.py`), while `publish_data()` additionally withholds the entity for `DISCHARGE_TARGET_UNSUPPORTED_MODELS`/`_TYPES` — on an unsupported-model fleet the arg points at `number.predbat_givtcp_N_discharge_target_soc` entities that are never created; benign so far (inverter.py leaves an unreadable discharge target alone, `No current discharge target to read`), fix shape = consult the same unsupported-model set or `published_discovery` like the discovery keys do. **GH#5209 (fixed in PR #5216; mechanism kept for pre-#5216 logs):** pre-#5216 `automatic_config()` filled every per-inverter arg list **positionally** from `discovered`, so an endpoint down at the discovery pass shifted every later endpoint down one slot, and `rediscover()`'s deliberate append-only recovery turned the shift into a rotation on recovery — recovery did not heal. Writes route by the index embedded in the entity id (`_parse_entity`), so the misroute reaches hardware; log signature `Inverter 1 current charge rate is 3000W and new target is 2600W` paired with `GivTCP: Inverter 2 set charge rate 2600 via REST successful` alternating each cycle, plus `battery_calibration ... has 2 entries, expected 3` / `Out of range index 2` for claimed-but-user-unconfigured keys (`_keep_configured_tail()` returned the short list unchanged when `_configured_value()` was None — only those keys warned while the user's own apps.yaml lists got their tail preserved). Debug-yaml diagnosis: any `predbat_givtcp_<n>_*` entry whose list position is not `<n>`. Workarounds at any version: restart Predbat once every endpoint answers, or `givtcp_automatic: False`. Design note on the fix: at startup a dead leading endpoint cannot be told apart from a leading placeholder URL, and #5216 chooses identity-by-index — `givtcp_rest: [dead, live]` becomes two inverters, not one (`num_inverters` cannot arbitrate; the GivTCP template tells REST users to delete it). **One design rule left by #5216's own development:** during the PR's cleanup it was found that the re-probe adoption pass judged every `published_discovery` gate before the adopted endpoint had published anything — `run()` calls `publish_data()`, then `rediscover()` appends the newly answering index, then `automatic_config()` re-runs because `discovered != configured_for`, so each `all(key in self.published_discovery.get(n, set()) for n in discovered)` gate failed on the adoption cycle and discovery-gated keys (`soc_max`, `battery_calibration`, `inverter_limit`, …) were skipped or handed back while the gates reading `rest_data` directly (`rest_v3`, `pause_mode_supported`, …) judged correctly. The merged PR re-publishes after a fleet-growing `rediscover()` (`givtcp.py`, the `if rediscover:` block) and adds a claim hand-back (`self.claimed_from`), and that ordering is the rule for anyone touching the path: **any gate over `published_discovery` must not be judged in the cycle where `rediscover()` appended an index.** On pre-#5216 builds (no hand-back, no wait) a *tail* endpoint adopted late never has its discovery keys claimed at all — suspected, unverified against a build without the PR — so a "late GivTCP inverter shows 8 kWh / no calibration sensor until restart" report on an older build maps here. **A missing inverter-details block silently stales the published sensors and fakes a clock-skew alarm (GH#5334, code-read on main 57ec7bf1):** `inverter_details()` (givtcp_rest.py) has no log and returns `{}` when `Invertor_Details` is empty (normal on v3) and `raw.invertor.serial_number` is null, which is exactly the Gateway/EMS shape upstream omits from `/readData` (britkat1980/giv_tcp#597 is the upstream fix, unmerged at triage time); `publish_data()` then gates each detail-block sensor on presence — `inverter_time`, `soc_max`, `battery_rate_max`, `inverter_limit` — so HA keeps the entity's *last* state frozen and stale looks identical to live, while `serial_number` itself is published unconditionally: **the discriminating tell is the serial sensor reading unknown while inverter_time stays frozen**. Predbat reads that frozen `inverter_time` and `check_clock_skew()` (inverter.py) crosses the ≥30-min restart threshold every cycle — repeating `Warn: Inverter time is <old>, Predbat computer time <now>, this is <growing> minutes skewed` and `Warn: Inverter control auto restart trigger: Clock skew >= 30 minutes command []` (empty `command` = no `auto_restart` configured) with the inverter clock actually fine — not the clock, not HA time sync. Fix shapes discussed, maintainer's call: a per-cycle warn when the details block resolves empty; a fallback lookup over top-level dict blocks for `Invertor_Serial_Number`/`Invertor_Time` (unambiguous only at exactly one candidate); or the cheapest anti-misdiagnosis — publish an explicit `unknown`/`unavailable` `inverter_time` instead of skipping, which `Inverter` already treats as "no reading" and skips skew detection for without a restart trigger. Any "Predbat-published entity looks frozen but claims to be live" report on a REST-backed component is a candidate for this same `if value:` publish-gate class. **GH#5386 (2026-10-04, open) — a read-back miss larger than the REST verify tolerance never converges:** the REST verify compares within `write_tolerance_watts()/12` (charge) and `/25` (discharge), where `write_tolerance_watts()` is the max rate GivTCP itself reports (`Invertor_Max_Bat_Rate`, 2600 W fallback); the reporter's 6000 W write read back 5100 while GivTCP's own log said the POST succeeded, so every ~2s retry and every cycle re-warns, and the entity-level deadband (`rate_tolerances()`'s one-step `fuzzy_below`) cannot cover a ~9-step miss either. Predbat's REST mirror entities **are** the read-back, so the debug yaml alone (mirror state vs the target) proves non-convergence without GivTCP logs. Why the register sits at 5100 is open — a stale cached Control read vs a firmware/BMS cap (the mirror's own max attribute says 6000, inconsistent with a pure cap) — and the suspected fix shape (accept GivTCP's POST success, a Solis-style `verify_settle_seconds` re-read, or sizing the tolerance to the firmware cap) is untested. Family so far: #5324 (exactly one rounding step; fixed by PR #5348's `fuzzy_below`), #2882 (~1 kW miss, open), #5386 (900 W miss, open). | `inverter` |
 | Compare (`compare.py`) | `apply_hardware_overrides()` overrode four attributes and not `battery_rate_max_export`, which is the one the export prediction path actually uses (`prediction.py:906`), so an `override_battery_rate_max_discharge_kw` left force-export slots pinned at the real hardware rate - probe-verified against the reporter's debug yaml, fixed in PR #4897 with an explicit fifth key (GH#4895). `battery_rate_max_export` itself is `min(inverter_limit_export, battery_rate_max_raw)` (`inverter.py:431`), and the prediction still clamps export draw at the un-overridden `inverter_limit`, so a "model a bigger inverter" scenario needs `override_inverter_limit_kw` as well or it is inert. A negative "True cost" alongside zero import is **not** phantom export revenue: compare reports `metric_end - metric_start`, and `compute_metric()` credits the end-of-scenario battery at the *replacement import* rate (`plan.py:1769`), deliberately ignoring the export rate - so a scenario ending 25 kWh fuller on free PV scores negative by construction. Likewise a non-zero Export column under `rates_export: 0` is forced PV-clipping export, not a planned export slot; banning export outright is a missing feature (GH#2446), not a compare bug (GH#4881, confirmed twice, the second time against the reporter's own detailed plans). GH#2033 (bumped 2026-09-18, code-verified): the table's Import and Export columns are whole-horizon totals, not plan-slot sums — `run_prediction()` accumulates `import_kwh_battery`/`import_kwh_house`/`export_kwh` for every modelled minute of grid flow and seeds `export_kwh` from `export_today_now`, so PV export outside slots, scenario-plan arbitrage and today-so-far export all count, and a planned discharge slot exports regardless of load vs solar. Deliberate: compare/annual pass `include_manual_api=False` to `basic_rates()` so live manual_api overrides never leak into simulated tariffs, while a tariff's own `rates_import`/`rates_export` replaces the live rates wholesale (no `prev` passed) — per-tariff `rates_*_override` blocks are the workaround; and #2033's "identical friendly names" premise is contradicted by the code (`"Compare " + name`), so duplicate names likely come from duplicate `name:` values in `compare_list`. GH#5133 (fixed on **open PR #5143, unmerged** - keep the mechanism for logs on stock main): `publish_data()` built the entity id by plain concatenation with no slug sanitising, so a `compare_list` id containing `/` produced `POST /api/states/predbat.compare_tariff_IGO/Prime` - and that 404 is the route not matching, not a missing entity. The warns arrive **in pairs** - `set_state()` does the POST and then a read-back `update_state()` GET, and both warn - so two identical 404 warns per cycle anywhere on `/api/states/...` means one bad write, not two problems. And `comparisons.yaml` was never pruned against `compare_list`, so a renamed or deleted tariff id kept republishing for as long as the file survived: the bad id existed only in the persisted yaml (the reporter's current apps.yaml looked fine), so grepping the attached config proves nothing - ask for `comparisons.yaml`, delete the stale entry and restart. The PR adds `tariff_entity_slug()` and `prune_comparisons()`. From the same PR's review (still unmerged, so observations about stock main): the tree has four ad-hoc entity-id slugify helpers with different semantics - `gateway.py`'s `_ev_suffix` drops non-`[a-z0-9_]` characters, `sigenergy.py`'s `_system_slug` folds only `-`, web.py's `re.sub` neither collapses runs nor strips, the PR's `tariff_entity_slug` collapses and strips - and `dashboard_item()` validates nothing, so on any "invalid entity id / 404 on publish" report check which slugify helper built the id; none caps length, and HA rejects object ids over 64 characters. A history lookup that passes an unknown entity id returns `None` silently - a wasted round-trip, **not** a 404 - so don't accept "it 404s" about the history path. With `compare_list` commented out, stored tariffs are republished **once per Predbat start** (only `load_yaml` reaches `publish_data` without the `compare_list` early return), so the tell for "sensors for a tariff I deleted keep coming back" is a restart, not the 5-minute cycle; and stored compare results are looked up by the raw `compare_list` id everywhere (`run_all`'s prior-SoC carry-forward, `select_best`, web.py's Compare page), so a fix that normalises tariff identity must *move* the stored entry to the configured id - keeping the old key preserved nothing, because nothing looked it up under the new id. **GH#5180 (code-read on main, still live):** the /compare tab's `Existing` badge is broken by any event stamped into the live rate tables at the moment compare runs — `Compare.run_all()` snapshots its baseline from the live `self.pb.rate_import`/`rate_export` (`compare.py:598-599`), which by then carry the IO/saving/free/Axle stamps, while `fetch_rates()` refetches the compared tariff's plain rate card and never re-applies any event — so the minute compare differs at every event minute and `result["existing_tariff"]` goes False, withholding the badge (web.py) for that whole run. Compare fires daily at midnight or on the `compare_active` switch, so an event active at that moment is enough. Cost figures are unaffected: every scenario runs on its own refetched plain card. In-tree fix baselines already exist: `rate_import_no_io` (also excludes manual overrides — but the compared refetch never applies manual rates either, so override users already fail this comparison today; the new base makes that consistent rather than opening a gap) and `rate_export_base` (comment says saving sessions and overrides stay out of it). `test_compare` has zero setup for saving/free/IO slots and never asserts `existing_tariff` — a green compare test says nothing about the event-active path. Suspected, not verified: the `run_all()` finally block restores everything it saved **except** `rate_import`/`rate_export`, so the live tables may hold the last compared tariff's rates until the next `fetch_sensor_data()` rebuild — if "the live plan used a different tariff's rates right after running a comparison" is reported, start there. | `compare` |
-| Car charging (`plan.py`, `fetch.py`, `execute.py`) | `car_charge_slot_kwh()` returns the raw slot kWh with no limit awareness and feeds the HTML Car kWh column, the JSON plan, the status sentence and the `car_charging_slot` attributes, while `prediction.py:799` and `prediction_kernel.cpp:827` clamp unconditionally - so "car kWh is displayed but the cost never moves" is a display-vs-model disagreement, not a planning bug. The *rate* side of beyond-cap IOG energy is already handled (`car_charge_slot_rate`/`car_rate_premium`); only the volume side is untrimmed, and `octopus_intelligent_consider_full` (default False, expert-hidden) is the only thing that trims it (GH#4888). The manual car SoC is prediction-fed and never a measurement: `predbat.py` exposes `car_charging_soc_next`, which `prediction.py` sets from the modelled first-step car SoC, so it never resets on its own (GH#4889). `update_car_charging_power()` (`execute.py`) sums *every* entity listed under `car_charging_power` - that list describes chargers, not cars, so two per-car template sensors reading one shared charger are added together; the summing is confirmed, the reporter's doubled figure was not (GH#4879). Per-car rate config pre-#5383: `car_charging_rate` was read by suffixed name as `float(get_arg("car_charging_rate" + postfix))` (`fetch.py`), so an apps.yaml **list** raised `float(list)` — a `TypeError` every main-loop cycle while the car fell back to 7.4 kW — and a fresh item (the HA input_number not yet created) had `load_user_config()` seed `item["default"] = self.args[name]` with no type validation, letting the invalid seed survive (`userinterface.py`; a comma-string value such as `car_charging_rate: "11.0, 11.0"` hits the same `float()` — suspected, not reproduced). **Fixed in PR #5383 (merged 2026-10-04, unreleased at the time of writing, post-v9.3.5) — keep the pre-#5383 mechanism for older logs:** `car_list_arg()` (`userinterface.py`) now splits a base-key list across the per-car slots (`car_charging_rate: [11.0, 6.5]` → car 0 11.0 / car 1 6.5), a per-car suffixed scalar (`car_charging_rate_1: 3.6`) wins over its list slice, and a list on a per-car suffix is ignored in favour of the base list's slice (warned when it stands alone) — `test_fetch_config_options.py` pins all three. The list-vs-postfix asymmetry is closed: `car_charging_limit`/`car_charging_battery_size` had always accepted lists (read with `index=car_n`) and `car_charging_rate` now accepts them. The `car_charging_slots` refresh gate that checked only `car_charging_planned` now also checks `car_charging_now` (`fetch.py:1337`, GH#4795 / PR #4796). **GH#4967 is now fixed (PR #4971, commit `4a045e07`) - the mechanism is kept for anyone reading a pre-#4971 log.** The prediction used to clamp modelled car load at the real `car_charging_limit` unconditionally, via an identical line in both `prediction.py` and `prediction_kernel.cpp`, with the discharge hold below each gated on `car_load_scale > 0` - so once the modelled car filled part-way through a slot the hold released for the rest of the window, regardless of the `octopus_intelligent_consider_full` switch, which never reached `predict()` at all (its only live uses were the read, the `False` default and the gated trim inside `load_octopus_slots()`). The fix routes the prediction through a model-facing `car_charging_limit_model` instead: with `octopus_intelligent_consider_full` off (the default) cars carrying IOG slots get `CAR_CHARGING_LIMIT_UNCAPPED`, so the fill clamp is deliberately inert and the hold releases only when the modelled car actually fills, while the real `car_charging_limit` is left untouched for `execute.py`'s "car already charged" decision, `plan_car_charging()` and `load_octopus_slots()` (it is taken **by reference** by `Prediction`, which is why the fix assigns a separate value rather than mutating it). Both engines take the model limit from the same attribute so they cannot diverge here, `update_car_manual_soc()` caps the manual car SoC write-back at the real limit because the modelled SoC can now overshoot, and `annual.py`/`compare.py` reset the attribute so an override cannot leak into their re-planning; `test_multi_car_iog.py` now asserts the switch's effect on the prediction. One interaction to re-check when the GH#4952 Ohme `watts x hours` overestimate is fixed: with consider_full off the model limit is uncapped for IOG-slot cars, so the fill clamp no longer bounds modelled car load for them either. The History view's separate reconstruction of car slots from the car energy sensor had its own faked-clock bug — GH#5004, under "The shared clock" above, fixed in PR #5025 (v9.0.2). The "Hold for car" discharge hold has **two different gate conditions** (GH#5146): live execution (`execute.py`'s carHolding block) requires an active car slot with `kwh > 0`, the car not already at limit, and skips entirely while an export window is executing (`if not isExporting:`), while the plan model (`prediction.py`, `discharge_rate_now = battery_rate_min`) holds on `(car_load_scale > 0) and (not car_charging_from_battery) and set_charge_window` — no live state, no kwh check, and `set_charge_window` is the plan's charge-window flag, not "this minute is in one". So the plan can hold in slots the live code charges through and vice versa: verify a "plan says X, status says Y" report against the right gate — and since PR #5147 the plan-table display *does* take `hold_for_car` from the prediction's own record (`predict_car_hold_best`, shown for at least half a slot on Demand rows), so a display disagreement is a version question, not a gate question. The export-window exemption is deliberate and test-codified (`discharge_car_full_bat2` in `test_execute.py` asserts export + car slot + the switch off → Exporting, no pause), so it is not a #2380 regression; if the reporter has a working `car_charging_now` sensor the overlapping export window is dropped on the next replan, so *sustained* discharge means the now-slot was never created — on pre-#5245 builds a numeric power sensor read as "not charging", but since PR #5245/#5267 `car_charging_now` can *be* a charging power sensor (a number in watts counts as charging from `CAR_CHARGING_NOW_POWER_W`, `car_charging_now_value()` in `fetch.py`), so check the version before reading a power sensor as absent evidence. Read-only users cannot see the status sensor's "Hold for car" at all — `predbat.status` is forced to "Read-Only" first (`execute.py`). **"Hold for car" is the hold working as designed, and there is no native EV-SoC or price gate anywhere (GH#4572, read on main 57ec7bf1):** the hold fires when `set_charge_window` is on and `car_charging_from_battery` is off (config.py) whenever a car is charging now or inside a planned slot; it stops the battery *feeding the car* (pause mode, else rate 0) and never blocks the car charging from the grid, so "battery idle while the EV charges on PV-then-grid" is the intended steady state and the charger's start/stop is outside the hold's scope. The full car config surface (`car_charging_energy_scale`/`_threshold`/`_rate`/`_loss`/`_hold`, `car_energy_reported_load`, `_manual_soc(_kwh)`, `_plan_smart`, `_plan_max_price`, `_from_battery`, `_plan_time`) has **no parameter gating charging on the EV's SoC and no enforceable price threshold** — `car_charging_soc`/`car_charging_limit` in apps.yaml are the car's *input* sensors, not thresholds — so "charge the EV only when X" is an enhancement ask; the maintainer-side answer is external HA automation, and the in-tree tooling direction is open PR #3791 (`predbat.solar_surplus_power` = `min(max(0, grid + car_power - max(0, battery)), max(0, pv))` + `binary_sensor.predbat_force_export_slot`, sensor-only after being pared back, still unmerged with an owner review round outstanding 2026-10-02). Two slot-selection facts from October 2026: the smart car planner ranks slots by **import price only** — `plan_car_charging()` reads `window["average"]`, while each `low_rates` window already carries an `export` average (`rate_scan_window(..., alt_rates=self.rate_export)`, `fetch.py`) that only `plan_iboost_smart()` consumes — so a cheap import band overlapping daylight is mis-ranked exactly when the export price beats it, and extra car kWh served from surplus PV costs the band's export price, not its import price (GH#5384, enhancement). And a `re:` config arg matches the **first** matching entity only — `resolve_arg_re()` breaks on the first hit (`userinterface.py`) — so a `car_charging_now: re:(sensor.wallbox_portal_status_description|sensor.myenergi_zappi_[0-9a-z]+_plug_status)` alternation consults exactly one detection sensor and the other integration's is never read (GH#5390, dump-verified: the resolved args carried a single plug_status entity); an alternation does not combine sensors, name one entity per arg. **GE "Charging, Hold for car" stall at the reserve (GH#5419, verified from the reporter's own predbat.log/GivTCP write log/debug yaml, 2026-10-06):** the live hold writes `discharge_rate = 0` with no isCharging guard (execute.py's carHolding block — the reserve bump at the same branch deliberately sits out charging, #3899) — refs verified exact on main `82bee43e`, and`git diff v9.3.1..HEAD` leaves the block untouched, so not a regression. Fault state: charge window enabled + target/rate read back correct + reserve floor + **discharge rate 0 + SoC exactly at the reserve** — a GE inverter (GIV-HY3.6 via GivTCP REST) declines to grid-charge, and charges normally within ~1 min of the hold's rate-0 being restored every time; the plan manufactured the collision by exporting down to the reserve immediately before the hold began. Suspected, not verified: the inverter-side rule itself (GE firmware won't grid-charge at SoC ≤ reserve with discharge disabled; a firmware OTA cannot be ruled out). Fix direction flagged to the maintainer: skip the rate-0 hold while isCharging, symmetric to the reserve skip — residual risk: in a dead charge window the inverter keeps its normal discharge rate with only the reserve floor preventing battery→car feed, a written register on GE, not on every inverter type. | `octopus_*`, `car_charging` |
+| Car charging (`plan.py`, `fetch.py`, `execute.py`) | `car_charge_slot_kwh()` returns the raw slot kWh with no limit awareness and feeds the HTML Car kWh column, the JSON plan, the status sentence and the `car_charging_slot` attributes, while `prediction.py:799` and `prediction_kernel.cpp:827` clamp unconditionally - so "car kWh is displayed but the cost never moves" is a display-vs-model disagreement, not a planning bug. The *rate* side of beyond-cap IOG energy is already handled (`car_charge_slot_rate`/`car_rate_premium`); only the volume side is untrimmed, and `octopus_intelligent_consider_full` (default False, expert-hidden) is the only thing that trims it (GH#4888; #5403, v9.3.6: when on it also cuts future dispatch minutes no car can use, with a replan trigger logging `Octopus Intelligent: dispatch minutes no car needs changed to N` that must be grepped for before reaching for it on a churn report). The manual car SoC is prediction-fed and never a measurement: `predbat.py` exposes `car_charging_soc_next`, which `prediction.py` sets from the modelled first-step car SoC, so it never resets on its own (GH#4889). `update_car_charging_power()` (`execute.py`) sums *every* entity listed under `car_charging_power` - that list describes chargers, not cars, so two per-car template sensors reading one shared charger are added together; the summing is confirmed, the reporter's doubled figure was not (GH#4879). Per-car rate config pre-#5383: `car_charging_rate` was read by suffixed name as `float(get_arg("car_charging_rate" + postfix))` (`fetch.py`), so an apps.yaml **list** raised `float(list)` — a `TypeError` every main-loop cycle while the car fell back to 7.4 kW — and a fresh item (the HA input_number not yet created) had `load_user_config()` seed `item["default"] = self.args[name]` with no type validation, letting the invalid seed survive (`userinterface.py`; a comma-string value such as `car_charging_rate: "11.0, 11.0"` hits the same `float()` — suspected, not reproduced). **Fixed in PR #5383 (merged 2026-10-04, unreleased at the time of writing, post-v9.3.5) — keep the pre-#5383 mechanism for older logs:** `car_list_arg()` (`userinterface.py`) now splits a base-key list across the per-car slots (`car_charging_rate: [11.0, 6.5]` → car 0 11.0 / car 1 6.5), a per-car suffixed scalar (`car_charging_rate_1: 3.6`) wins over its list slice, and a list on a per-car suffix is ignored in favour of the base list's slice (warned when it stands alone) — `test_fetch_config_options.py` pins all three. The list-vs-postfix asymmetry is closed: `car_charging_limit`/`car_charging_battery_size` had always accepted lists (read with `index=car_n`) and `car_charging_rate` now accepts them. The `car_charging_slots` refresh gate that checked only `car_charging_planned` now also checks `car_charging_now` (`fetch.py:1337`, GH#4795 / PR #4796). **GH#4967 is now fixed (PR #4971, commit `4a045e07`) - the mechanism is kept for anyone reading a pre-#4971 log.** The prediction used to clamp modelled car load at the real `car_charging_limit` unconditionally, via an identical line in both `prediction.py` and `prediction_kernel.cpp`, with the discharge hold below each gated on `car_load_scale > 0` - so once the modelled car filled part-way through a slot the hold released for the rest of the window, regardless of the `octopus_intelligent_consider_full` switch, which never reached `predict()` at all (its only live uses were the read, the `False` default and the gated trim inside `load_octopus_slots()`). The fix routes the prediction through a model-facing `car_charging_limit_model` instead: with `octopus_intelligent_consider_full` off (the default) cars carrying IOG slots get `CAR_CHARGING_LIMIT_UNCAPPED`, so the fill clamp is deliberately inert and the hold releases only when the modelled car actually fills, while the real `car_charging_limit` is left untouched for `execute.py`'s "car already charged" decision, `plan_car_charging()` and `load_octopus_slots()` (it is taken **by reference** by `Prediction`, which is why the fix assigns a separate value rather than mutating it). Both engines take the model limit from the same attribute so they cannot diverge here, `update_car_manual_soc()` caps the manual car SoC write-back at the real limit because the modelled SoC can now overshoot, and `annual.py`/`compare.py` reset the attribute so an override cannot leak into their re-planning; `test_multi_car_iog.py` now asserts the switch's effect on the prediction. One interaction to re-check when the GH#4952 Ohme `watts x hours` overestimate is fixed: with consider_full off the model limit is uncapped for IOG-slot cars, so the fill clamp no longer bounds modelled car load for them either. The History view's separate reconstruction of car slots from the car energy sensor had its own faked-clock bug — GH#5004, under "The shared clock" above, fixed in PR #5025 (v9.0.2). The "Hold for car" discharge hold has **two different gate conditions** (GH#5146): live execution (`execute.py`'s carHolding block) requires an active car slot with `kwh > 0`, the car not already at limit, and skips entirely while an export window is executing (`if not isExporting:`), while the plan model (`prediction.py`, `discharge_rate_now = battery_rate_min`) holds on `(car_load_scale > 0) and (not car_charging_from_battery) and set_charge_window` — no live state, no kwh check, and `set_charge_window` is the plan's charge-window flag, not "this minute is in one". So the plan can hold in slots the live code charges through and vice versa: verify a "plan says X, status says Y" report against the right gate — and since PR #5147 the plan-table display *does* take `hold_for_car` from the prediction's own record (`predict_car_hold_best`, shown for at least half a slot on Demand rows), so a display disagreement is a version question, not a gate question. The export-window exemption is deliberate and test-codified (`discharge_car_full_bat2` in `test_execute.py` asserts export + car slot + the switch off → Exporting, no pause), so it is not a #2380 regression; if the reporter has a working `car_charging_now` sensor the overlapping export window is dropped on the next replan, so *sustained* discharge means the now-slot was never created — on pre-#5245 builds a numeric power sensor read as "not charging", but since PR #5245/#5267 `car_charging_now` can *be* a charging power sensor (a number in watts counts as charging from `CAR_CHARGING_NOW_POWER_W`, `car_charging_now_value()` in `fetch.py`), so check the version before reading a power sensor as absent evidence. Read-only users cannot see the status sensor's "Hold for car" at all — `predbat.status` is forced to "Read-Only" first (`execute.py`). **"Hold for car" is the hold working as designed, and there is no native EV-SoC or price gate anywhere (GH#4572, read on main 57ec7bf1):** the hold fires when `set_charge_window` is on and `car_charging_from_battery` is off (config.py) whenever a car is charging now or inside a planned slot; it stops the battery *feeding the car* (pause mode, else rate 0) and never blocks the car charging from the grid, so "battery idle while the EV charges on PV-then-grid" is the intended steady state and the charger's start/stop is outside the hold's scope. The full car config surface (`car_charging_energy_scale`/`_threshold`/`_rate`/`_loss`/`_hold`, `car_energy_reported_load`, `_manual_soc(_kwh)`, `_plan_smart`, `_plan_max_price`, `_from_battery`, `_plan_time`) has **no parameter gating charging on the EV's SoC and no enforceable price threshold** — `car_charging_soc`/`car_charging_limit` in apps.yaml are the car's *input* sensors, not thresholds — so "charge the EV only when X" is an enhancement ask; the maintainer-side answer is external HA automation, and the in-tree tooling direction is open PR #3791 (`predbat.solar_surplus_power` = `min(max(0, grid + car_power - max(0, battery)), max(0, pv))` + `binary_sensor.predbat_force_export_slot`, sensor-only after being pared back, still unmerged with an owner review round outstanding 2026-10-02). Two slot-selection facts from October 2026: the smart car planner ranks slots by **import price only** — `plan_car_charging()` reads `window["average"]`, while each `low_rates` window already carries an `export` average (`rate_scan_window(..., alt_rates=self.rate_export)`, `fetch.py`) that only `plan_iboost_smart()` consumes — so a cheap import band overlapping daylight is mis-ranked exactly when the export price beats it, and extra car kWh served from surplus PV costs the band's export price, not its import price (GH#5384, enhancement). And a `re:` config arg matches the **first** matching entity only — `resolve_arg_re()` breaks on the first hit (`userinterface.py`) — so a `car_charging_now: re:(sensor.wallbox_portal_status_description|sensor.myenergi_zappi_[0-9a-z]+_plug_status)` alternation consults exactly one detection sensor and the other integration's is never read (GH#5390, dump-verified: the resolved args carried a single plug_status entity); an alternation does not combine sensors, name one entity per arg. | `octopus_*`, `car_charging` |
 | Web, MCP and Chat (`web.py`, `web_mcp.py`, `chat.py`) | `html_plan_override` responds `{"success": true}` unconditionally and discards the override write's result, and the `run_in_executor()` it goes through submits to a thread pool without ever calling `.result()` on the future (`hass.py`), so an exception while persisting a slot override is swallowed with no log line anywhere. That code path is confirmed by reading; its link to the "manual override partially ignored" report it was found under is **not** (GH#3078). The MCP OAuth endpoint advertises `client_secret_basic`, but `oauth_token`/`_handle_authorization_code` only ever read the POST body and never the `Authorization` header, so a client authenticating via HTTP Basic looks like it is missing `client_id`; and `oauth_metadata_mcp` sets `issuer` to the bare host rather than `{base_url}/mcp`, which RFC 8414 requires when the metadata is served under a path suffix (GH#4799). Chat's provider table (`PROVIDERS`, `chat.py`) has no first-class Gemini entry; `type: openai` against Google's OpenAI-compatible endpoint works for plain chat but not for tool calling, because the tool-call accumulator keys fragments solely on `fragment.get("index", 0)`, so index-less SSE fragments collapse every parallel call into one garbled slot - the sibling `reasoning_details` accumulator in the same function already handles index-less fragments - and `thought_signature` appears nowhere in `chat.py` while the assistant message is replayed verbatim, so Gemini's required signature echo-back is dropped (GH#4904, both confirmed by reading, neither tested against the live API). Also `ChatRequestError.friendly()` hardcodes OpenRouter wording as its generic fallback (`OpenRouter returned HTTP N`), and so do the 402 and 429 branches — for any provider a non-401 HTTP failure is surfaced to the user as an OpenRouter error, confirmed live in GH#5074 where a direct-OpenAI 400 reached the reporter as "OpenRouter returned HTTP 400"; read the reporter's provider entry in apps.yaml, not the message. `_stream_chunks()` always appends `/chat/completions` to the user-configurable base URL and the catalogue fetch `/models`, so there is no config-only route to another path. The header status icons read **published entities, not live power**: `get_battery_status_icon()` (`web.py`) builds the SoC+icons string from `predbat.soc_kw` and the states of `binary_sensor.<prefix>_charging`/`_exporting`, which `set_charge_export_status()` (`output.py`) publishes from the end of `execute_plan()` — there is no path from the icon code back to the plan, so anything the header should distinguish has to be published by execute first (GH#5125/PR #5131). One string, three render sites — the page header, the dash Status table SoC row, and `html_api_get_status`'s `battery_html` JSON — all through `get_header_html`; page CSS lives in `get_header_html()` in `web_helper.py`, not a stylesheet, and the established pattern for anything coloured is a base rule plus a `body.dark-mode .foo` override, so an inline `style=` on generated markup cannot follow dark mode. **The chat model picker's "Default (…)" row was a silent no-op after any manual pick (GH#5230, fixed in PR #5274, merged 2026-09-27 — keep the mechanism for pre-#5274 logs):** `html_chat_model()` (`web_chat.py`) used to treat the empty id the Default row posts as "clear the conversation override" but called `set_selected_model()` only behind `if model_id:` — so the per-provider *remembered* selection (which outranks the apps.yaml default in `resolve_model()`'s chain: conversation override → remembered selection → default) kept winning in both the server and the browser's `effectiveModel()` mirror, and Default was unreachable. Editing the provider's `model:` in apps.yaml did **not** work around it (the remembered choice outranks it); clearing Predbat's chat storage did. #5274 calls `set_selected_model()` unconditionally (`chat_store.py` already pops the entry on a falsy id, so the default is not pinned) and clears the JS mirror's remembered copy before the picker is redrawn, and the new tests finally cover the empty-id Default path. Two /apps editor traps (PR #5243 review, code-read on main — the PR itself merged as 6d394b7a): `resolve_arg()` resolves templates with `value.format(**self.args)` and catches **only `KeyError`**, so an unbalanced brace (`abc{def`) raises `ValueError` and a positional `{0}` raises `IndexError` — `WebInterface.resolve_value_raw()` calls it for every string containing `{` on the `/apps` page render path, so a non-credential value like that is *suspected, not reproduced* to 500 the whole page (credentials escape on the page only because `mask_secret_args()` has already turned them into `xxx`); and the editor addresses nested values by dotted path (`parse_yaml_path()` splits on `.` and `[n]`) while `render_type()` builds `data-nested-path` from the literal key, so a dict key that literally contains `.` or `[` cannot be edited correctly — the save targets the wrong node or raises. The quote-bearing-value truncation (`value="${currentValue}"` in the JS) that #5243 fixed — input values now set via `input.value` and `escapeHtml` — matters only on pre-#5243 trees. **Plugins: endpoints work, nav is hardcoded (GH#5367, code-verified on main).** Plugin `on_web_start` hooks fire in `WebInterface.start()` (`web.py`) *before* the `registered_endpoints` loop mounts routes — the plugin system is initialised early (`predbat.py`) exactly for that — so a plugin page that 404s is an `on_web_start` problem, not a routing one. But `get_header_html()` (`web_helper.py`) hardcodes the nav menu with no plugin extension point, and `registered_endpoints` is consumed only by `start()` — no page lists them — so a plugin page is reachable only by typing its URL, despite `docs/plugins.md` promising "additional web interfaces and dashboards". "My plugin's page loads but I can't get to it from the Predbat UI" → the gap is nav/UI exposure, not the endpoint mechanism; fix shape is an opt-in plugin nav API (label+path, not auto-listing `registered_endpoints` — plugins register API-only paths too), matching the existing conditional-nav precedents (the Chat link's gate, "Metrics"). | `web_*`, `web_mcp`, `web_chat` |
 | Self-update (`github.py`, `download.py`) | The whole update path calls the GitHub REST API **unauthenticated** - the release check (`github.py:126`) and the file listing (`download.py:79`) - so it shares the 60-requests-per-hour-per-IP unauthenticated quota with everything else behind that egress IP. Behind CGNAT, a VPN or a shared proxy that 403s even though Predbat's own volume is one check per two hours; failures are deliberately never cached, so it then retries every 5-minute cycle. The listing request's error reaches only the addon stdout via `print()`, not the HA log a user would paste, which is why the report reads as "it just won't update" (GH#4886). A phone-hotspot A/B is the cheap differential test. | none |
 | Cloud / divergence modelling (`fetch.py`, `plan.py`) | Until v8.55.0 `step_data_history()`'s modulation was inert for every caller but one, through operator precedence: `int(...) + 1 if flip else 0` parses as `(int(...) + 1) if flip else 0`, a constant `0` whenever `flip` is False - and `plan.py`'s PV10 call was the sole `flip=True` caller. So `metric_cloud_enable` only ever perturbed the PV10 scenario, and `metric_load_divergence` - genuinely computed from load std-dev/mean in `output.py` - did nothing at all, on defaults-on settings, for as long as the block existed (GH#4870, evaluated directly rather than inferred; fixed in `f6c925d8` and then reworked again by the envelope model in `79f28f8f`). Worth knowing when reading a pre-v8.55.0 log or replaying an older debug dump: neither knob can be the explanation there. The review of the fix also found that switching the block *on* is the risky half - an empty `pv_forecast_minute10` pins `metric_cloud_coverage` at exactly 0.5 regardless of the weather, which is every non-Solcast user. |
@@ -160,7 +160,7 @@ Grep for the named symbol rather than trusting a line number.
 | `Warn: Unable to load the load forecast ... predheat.heat_energy$external` every cycle, startup validation `Invalid value None in element predheat.heat_energy$external`, and no `Predheat:` log lines despite the enable switch being on | Predheat's settings are not where the loader looks. `predheat_enable` is a Predbat UI switch, but the `predheat:` config domain must be nested **inside `pred_bat:`** — a top-level `predheat:` app section is silently discarded by the standalone loader, which reads only `config["pred_bat"]` (`hass.py:322`), and under AppDaemon a top-level section cannot load the bundled class at all. The whole `templates/predheat.yaml` is nested under `predbat:`; `docs/predheat.md` says "append it to the end of apps.yaml", which invites the mistake. `self.args.get("predheat", {})` then stays falsy so `PredHeat` is never instantiated (`predbat.py:2009`), the `predheat.heat_energy` entity is never published, and `fetch_extra_load_forecast` warns every cycle. Applies to any bundled sub-app configured by domain (the domain read is `self.args.get(domain, {}).get(arg, default)`) (GH#5002). |
 | Plan tab blank / no plan shown in HA | `predbat.cost_today` missing from HA recorder history — a recorder configuration problem (GH#3936). Test: `faq_recorder_config`, and see `docs/faq.md`. |
 | Exports happen at the wrong time | `plan.py` export optimisation order, plus `combine_export_slots` and `metric_min_improvement_export` in their config. |
-| IOG: battery flat inside an evening dispatch that runs on into 23:30, plan waits until 01:00 to charge | Two causes, both **fixed on main in PR #5304 (merged 2026-09-30, v9.3.3)** - keep the mechanisms for pre-#5304 logs. (1) Dispatches applied by `rate_add_io_slots()` (Predbat's own Octopus component) never set `io_adjusted`, so the PV10 "slot goes away" re-pricing never fired; it now flags the minutes it lowers, matching BottlecapDave's `adjust_intelligent_rates()`, which flags only rates it lowered - never the fixed off-peak or over-cap time. (2) With `car_charging_from_battery: false` the PV10 case still held the battery for the car in a vanished dispatch, so an early top-up had no value; PV10 now keeps the car at the dispatch rate but off the battery and releases the hold (kernel parity 15). Pre-fix check: `io_adjusted: {}` in the dump alongside live `octopus_slots`. The debug replay reuses the dump's `rate_import`/`io_adjusted` and never re-runs `rate_add_io_slots()`, even with `--redo`, so it cannot show a rate-overlay change - re-apply the overlay from `rate_import_no_io` in a wrapper. Also: with only the first (trusted) half hour left before 23:30 the benefit is under `metric_min_improvement_plan` (2p), so a dump taken late in the dispatch rightly shows no change. |
+| IOG: battery flat inside an evening dispatch that runs on into 23:30, plan waits until 01:00 to charge | Two causes, both **fixed on main in PR #5304 (merged 2026-09-30, v9.3.3)** - keep the mechanisms for pre-#5304 logs. (1) Dispatches applied by `rate_add_io_slots()` (Predbat's own Octopus component) never set `io_adjusted`, so the PV10 "slot goes away" re-pricing never fired; it now flags the minutes it lowers, matching BottlecapDave's `adjust_intelligent_rates()`, which flags only rates it lowered - never the fixed off-peak or over-cap time; and from PR #5392 (v9.3.6) "lowered" means by more than `IO_RATE_TOLERANCE` (a penny), so sub-penny dp2-rounding differences no longer mark the fixed off-peak window as adjusted or re-price it at the dispatch price. (2) With `car_charging_from_battery: false` the PV10 case still held the battery for the car in a vanished dispatch, so an early top-up had no value; PV10 now keeps the car at the dispatch rate but off the battery and releases the hold (kernel parity 15). Pre-fix check: `io_adjusted: {}` in the dump alongside live `octopus_slots`. The debug replay reuses the dump's `rate_import`/`io_adjusted` and never re-runs `rate_add_io_slots()`, even with `--redo`, so it cannot show a rate-overlay change - re-apply the overlay from `rate_import_no_io` in a wrapper. Also: with only the first (trusted) half hour left before 23:30 the benefit is under `metric_min_improvement_plan` (2p), so a dump taken late in the dispatch rightly shows no change. |
 | IOG: the house battery charges at the back of (or not at all during) a long smart-charge dispatch whose rate is equally cheap throughout | Within-block placement is **metric-neutral** — the cost model has no term preferring the earlier minutes of an equally-priced dispatch (GH#5330, probe-verified on the reporter's own replay at main 57ec7bf1: charged-early 10:00–11:30 = 148.5602p, continuous 10:00–12:00 = 148.5586p, the reporter's 10:30 late ramp = 148.5745p — equal to rounding), so placement is decided by search convergence plus plan retention (`metric_min_improvement_plan`) — **second live instance GH#5344** (GE Hybrid + one car on IOG, 2026-10-02, reporter-log-verified): the mid-band charge limit flapped between equally-scored shapes tick to tick and the overnight limit shrank toward a later cap (91% at 00:05 → 99% by 04:45 → 100%) as the horizon narrowed, which is the tie surface shrinking, not an execute fault; concurrent battery charging during the car's dispatch is allowed and executed on this shape, so do not read a car dispatch as a charge window blocker; the IOG earlier-charge skew that *does* exist (`IO_ADJUST_SLOPE`/`_PIVOT`/`_MAX_DISCOUNT`/`_MAX_PENALTY`/`_DISCOUNT_HORIZON`, `_io_rate_adjustment()` inside `sort_window_by_price_combined()`, plan.py) is ranking-level only — a signed gradient over `io_adjusted`-flagged window starts, reordering candidate windows, never a tie-break in the metric — so equal-cost placements land wherever the search converged. Suspected, not verified: PR #5304's PV10 "dispatch gone" hedge can *systematically* steer the planner to the un-flagged tail of a dispatch (the integration's `io_adjusted` flags can stop short of the dispatch end); and which search/retention path resolves equal-metric placements late. Cheap A/B technique: replay the dump, then compare direct `run_prediction()` variants — any placement fix must survive metric evaluation (a real cost term, or an explicit equal-metric tie-break that slides windows earlier inside an `io_adjusted` run). See also the previous row — pre-#5304 the dispatch minutes also were never `io_adjusted`-flagged, so the PV10 hedge did not even see them. |
 | IOG: the house battery charges cheaply for the first few minutes of a car dispatch that started the moment it was published, then the plan reverts | The car-need gate from PR #5403 (v9.3.6, for #4482) can only withhold dispatch minutes from the half hour **after** the current one — `octopus_surplus_from()` returns `(minutes_now // 30 + 1) * 30` (octopus.py; the docstring states the carve-out as deliberate: the car may already have charged in the current one), and `rate_add_io_slots()` skips only `minute in self.octopus_surplus` — so a same-instant-start dispatch still cheap-stamps the current half hour even with `octopus_intelligent_consider_full` on (GH#5424, diffed the published v9.3.5 tarball against main). The only veto inside that half hour is the dynamic-load evidence path — 3 min start + 2 min sensor grace (`DYNAMIC_LOAD_CAR_START_MINUTES`/`DYNAMIC_LOAD_CAR_SENSOR_MINUTES`, const.py) then replan + strip — and with `octopus_intelligent_trust_slots` off the running decision `dynamic_load_car_cancelled.get(car_n, default_cancelled)` (plan.py) can retain a stale not-cancelled from a dormant (slot-less) phase, so a trust-off user's first cycle inside a brand-new dispatch is still trusted until the grace evidence lands. The car-side skip (`Car 0 is already charged, ignoring additional charging slot`, execute.py) does not stop the house plan. So "upgrade fixes it" is not an answer to a same-instant-start dispatch report without checking which half hour the dispatch started in. Overnight dispatches inside the fixed 23:30–05:30 band are tariff-cheap regardless (`in_iog_fixed_window` excludes them from the gate) — don't attribute overnight dispatch-stamp plans to this class; #4516 is the open umbrella for the residual. |
 | "Random" exports that lose money on a flat export tariff | `metric_battery_value_scaling` below 1.0, then `metric_cloud_enable` being off. See "Check configuration before code" above (GH#4840) — the protective margin is ~1%, so a small reduction is enough. |
@@ -168,7 +168,8 @@ Grep for the named symbol rather than trusting a line number.
 | Battery pinned at 100% while the house imports at peak | Rule the plan out from the log before opening `plan.py`. `Completed run status <mode>` says what Predbat asked for and `Inverter N count register writes` says whether it changed anything: `Demand` on every cycle, zero writes, and a flat SoC together mean Predbat asked for discharge and something outside it refused - look at the reserve/min-SoC entity binding for that integration. Three successive analyses of GH#4961 theorised about a charge window spanning the peak that the reporter's own log shows never existed. |
 | Battery charges to 100% at Predbat start, on a `pred_bat_mode` change or on a `set_read_only` change (an inverter whose target register carries a floor, e.g. Victron ESS via `inverter_type: SK` or any profile with `has_charge_enable_time: False`) | `PredBat.reset_inverter()` (`execute.py`): "safe mode" writes `adjust_battery_target(100.0, False)` with no `inv_has_charge_enable_time` consult - on such a profile the target register carries the ESS minimum SoC, so 100% is a charge-to-100 order, not an idle value. Fire conditions: the 13 `reset_inverter: True` CONFIG items set `inverter_needs_reset` on any value change *including* None→value at startup (`userinterface.py`), and the two `reset_inverter_force` items (mode select, `set_read_only`) additionally set the force name on a user event - the forced paths run the reset even in Read-Only, where no further writes happen afterwards (the head of `reset_inverter()` gates on it, and `execute_plan()` skips the loop). Normally the 100 is corrected later the same `execute_plan()` cycle (`adjust_battery_target_multi(inverter, 0, ...)` for `has_charge_enable_time: False` inverters outside a window), so it stands only when nothing competes: entering Read-Only, or Predbat stopping right after the reset. Other 100% writes: the calibration branch (`execute.py`) and `self_test()` (only under the `INVERTER_TEST` dev constant). Fixture trap: the identically named `reset_inverter(my_predbat)` in `tests/test_infra.py` resets test state and is unrelated - grep-based test-coverage checks get fooled, and no test module covers the real write path. Docs trap: `docs/inverter-setup.md` links `templates/victron.yaml`, which does not exist (404, confirmed against the raw URL; the Victron route exists in prose and GH#2846's discussion only) - a "user following the docs" Victron report is following the issue discussion. | `inverter` |
 | `Hold charging target X%-Y%` while the house still imports at a peak price, or a peak-price charge window in the plan | Three code paths, all verified by reading on main (GH#2716; the family covers #1690 and #5083, and the maintainer marked the original #2716 case fixed in Nov 2025): (1) *plan side* — the high-price gate (`plan.py`, `price_key > best_price_charge_level`) guards only windows whose limit is still 0, so a window already carrying a limit from the levelling pass is exempt and `discard_unused_charge_slots()` keeps any `limit > 0` slot with no price re-check — a peak-price charge window in the plan is not by itself gate evidence, run the config-before-code checks first; (2) *"Hold charging" can be an armed-window fallback, not a hold* — the window-disabling hold needs `set_reserve_hold` + `set_reserve_enable` + `reserve_max >= target` or `inv_has_timed_pause`, and without either the fallback re-arms the charge window while the status still reads "Hold charging"; (3) *the pre-window freeze write* — a freeze-flavoured window inside `set_soc_minutes` before charging starts logs `not yet in charge freeze, holding target SoC at 100%` and deliberately writes a 100% target, so a 100% slot visible in the inverter portal during a displayed hold is that write, not a rogue one. Forensic decoder: the status suffix in `Hold charging target X%-Y%` is the **SoC–target pair**, not window start–end; and while charging the executor writes `max(charge_limit_best[0], reserve)` as the target, so SoC rising past the target means the writes never landed or the inverter ignores mid-window limit reductions (unresolved — needs the brand). | `inverter`, `plan` |
-| Predbat charged the car at peak rate, or planned a small house-battery top-up inside a peak window, with car charging enabled | Non-smart car planning is price-blind by design: **`car_charging_plan_smart` defaulted false until PR #5251 (GH#5237, merged 2026-09-26, released in v9.3.0) flipped the default to true** — `car_charging_plan_max_price` defaults 0 (never filters) either way, so pre-#5251 the car filled from the earliest low-rate-scan window onward regardless of price, and a plug-in mid-peak charged at peak (GH#2716, log-confirmed; smart mode is what re-sorts by price and honours the max-price skip, `plan.py`). On a #5251+ build a peak-rate car charge is no longer the default path — check the version before citing this explanation. Two more non-bug shapes from the same report: a saving-session reward is stamped onto the **import** rate too, which can push that hour above the low-rate scan threshold (`Rate thresholds ... import = rate_max - 0.5`), so the hour vanishes from `low_rates` and the car slot list shows a gap there — do not read that gap as a planning bug; and the in-progress peak-window house-battery top-up is the optimiser's keep response (`car_charging_from_battery: true` + `best_soc_keep` > 0 + the end-of-record SoC credit at replacement cost), same family as the "Hold charging target" row above — the high-price gate only skips windows whose limit is still 0, so a deliberately-selected window is exempt by design. Traps: `best_soc_keep` can be time-varying (automation or user), so the incident log and a later debug dump can disagree (4.0 vs 5.0 observed); and the `Calculate Best options` log line carries the plan-relevant config (`best_soc_keep`, `metric_battery_value_scaling`, ...) so you often don't need the yaml. | `car_charging` |
+| Plan table shows `HoldChrg -> N%` (and the status text "charging to N%") while the slot is actually charging to a higher limit | The State cell and the status text both classify from `charge_window_best[n]["target"]` — `calc_percent_limit(target, ...)` in `publish_html_plan()` (output.py), `HoldChrg` when it lands at/below the window's minimum predicted SoC — but `discard_unused_charge_slots()` only stamps `target = limit` when predicted SoC sits at/above the limit at both window start and end, and **otherwise `target` = predicted end-of-window SoC**: the rate-limited "charge can't complete" case. So a charge that cannot complete renders as `HoldChrg -> <predicted SoC>` and `get_charge_export_text()` says "charging to N%", while a genuine hold (`target == limit`) still shows HoldChrg. Freeze classification already reads the limit (`is_freeze_charge(limit)`), so the two classifications can also disagree. A rising charge keeps span-min below end-SoC, so completing charges label correctly *by accident* — a green `Chrg` in a quick manual check proves nothing about the classification input. GH#4445 (split from #4440, whose 4-Aug log is the live evidence; #4446 sibling fixed and closed), verified on main 2026-10-07 by reading + a scratch unit test on the publish_html_plan path, unmerged. Fixture gap: `test_plan_why_reason.py` builds charge windows **without** a `"target"` key, so only the `window.get("target", ...)` fallback path is exercised — pin it with windows whose `target` diverges from `charge_limit_best`. Fix shape: classify from the limit (the same percent `calc_percent_limit(limit, ...)`, one token at each site) and keep the predicted end-SoC as the debug parenthesis; that makes Chrg/HoldChrg consistent with the freeze check. |
+| Predbat charged the car at peak rate, or planned a small house-battery top-up inside a peak window, with car charging enabled | Non-smart car planning is price-blind by design: **`car_charging_plan_smart` defaulted false until PR #5251 (GH#5237, merged 2026-09-26, unreleased at the time of writing) flipped the default to true** — `car_charging_plan_max_price` defaults 0 (never filters) either way, so pre-#5251 the car filled from the earliest low-rate-scan window onward regardless of price, and a plug-in mid-peak charged at peak (GH#2716, log-confirmed; smart mode is what re-sorts by price and honours the max-price skip, `plan.py`). On a #5251+ build a peak-rate car charge is no longer the default path — check the version before citing this explanation. Two more non-bug shapes from the same report: a saving-session reward is stamped onto the **import** rate too, which can push that hour above the low-rate scan threshold (`Rate thresholds ... import = rate_max - 0.5`), so the hour vanishes from `low_rates` and the car slot list shows a gap there — do not read that gap as a planning bug; and the in-progress peak-window house-battery top-up is the optimiser's keep response (`car_charging_from_battery: true` + `best_soc_keep` > 0 + the end-of-record SoC credit at replacement cost), same family as the "Hold charging target" row above — the high-price gate only skips windows whose limit is still 0, so a deliberately-selected window is exempt by design. Traps: `best_soc_keep` can be time-varying (automation or user), so the incident log and a later debug dump can disagree (4.0 vs 5.0 observed); and the `Calculate Best options` log line carries the plan-relevant config (`best_soc_keep`, `metric_battery_value_scaling`, ...) so you often don't need the yaml. | `car_charging` |
 | Battery sits idle ("Hold for car") while the car charges from the grid on PV-then-grid — "why isn't the battery covering the house during the car's slot?" | Expected, not a bug: the hold (`execute.py`'s carHolding block) fires on `set_charge_window` with `car_charging_from_battery` off whenever a car is charging now or inside a planned slot, and it only stops the *battery feeding the car* — the car's grid charging and the charger's start/stop are outside its scope (GH#4572). "Charge the EV only when X" has no native knob at all: no EV-SoC gate, and the price knobs shape which slots are planned, not the charger — see the car-charging row above and open PR #3791 for the tooling direction. |
 | "Why does Predbat discharge to the house during cheap hours?" / low-rate hold requests | A structural model gap, not an optimiser mistake (GH#5122): any minute outside a selected charge AND export window falls into the `ECO Mode` branch of `predict()` (`prediction.py`), where the battery is modelled supplying whatever house load PV cannot cover — there is **no modelled action meaning "import from grid instead of discharging the battery"** outside a selected window. The only zero-discharge-outside-window modelled states are EV slots (`car_charging_from_battery: false`) and iBoost (`iboost_prevent_discharge`); all freeze/hold config is scoped to selected windows or manual periods. Also: a charge window covering a cheap gap saves only the round-trip loss on the cycled kWh — often ~2p, right at the default `metric_min_improvement_plan` gate of 2.0p — so a marginal window can legitimately fail to clear it (analysis, not verified against a real plan). | none |
 | "Predbat reports a GivEnergy error but I use Solis / Fox / GE-Cloud" | `inverter_type` was never set. It defaults to `GE` (`inverter.py`'s `get_arg("inverter_type", "GE")`, warning to the log only) and component auto-config is what actually writes it — roughly a dozen components (`solis.py`, `gecloud.py`, `fox.py`, `givtcp.py`, `sigenergy.py`, `gateway.py`, `teslemetry.py`, `deye.py`, `sunsynk.py`, `alphaess.py`, `solax.py`, `enphase.py`) — but auto-config only runs after discovery succeeds, while the component is registered as active the moment it is instantiated. So an outage or credential failure at startup leaves a non-GivEnergy inverter driven with GivEnergy capability flags, and every source read failure surfaces as `unable to read ... window - GivEnergy returned no data, check the GivEnergy credentials`, because `inverter_source_name()` prefers `INVERTER_DEF[inverter_type]["name"]` and `GE` is named "GivEnergy". Confirm from the debug yaml (no `inverter_type` key) plus the log line `Warn: ... inverter_type is not set in apps.yaml`, and treat any GE-specific behaviour in that log as assumed rather than configured. Self-heals on comms recovery; a manual `inverter_type` avoids it. The warning now names the assumed type and the component that should have set it — PR #4992 merged (`bdeb4e05`) after GH#4990, so on a current version the message reads `assuming <name> (<type>)` rather than a bare "assuming GE" (GH#4990). |
@@ -182,6 +183,7 @@ Grep for the named symbol rather than trusting a line number.
 | First cheap window charges to 0% / the charge target lands later than the cheap band | A metric tie in the charge-window SoC scan: the running-best adoption is non-strict `<=` (`plan.py`, `(metric + min_improvement_scaled) <= best_metric_first and metric <= best_metric`) and `min_improvement_scaled` gates only against the first (full-charge) candidate, so an exact tie lets the later, lower-SoC candidate replace the fuller one. Ties are realistic because `metric_battery_cycle` defaults 0, making charging 1 kWh cost ≈ its stored-energy credit (GH#5069). Neither knob fixes it: `metric_min_improvement` doesn't cover the running best, and raising `metric_battery_cycle` raises the stored-energy credit too. |
 | I set a config item above its maximum and it stays capped — no YAML workaround | An `APPS_SCHEMA` item's `max` is enforced as a hard clamp on **every** value source, not just the HA UI: the input_number branch of `userinterface.py` re-clamps after `float()` (deliberate — an apps.yaml override bypasses the HA entity's own min/max; `Warn: Config item ... clamping to ...` is the diagnostic). GH#5070 instance: `debug_history_count` max 50 while the capture path itself is uncapped (retention is `interval × count` in `_capture_debug_history()`), so widening the window was a schema change — PR #5071 (merged 2026-09-19) raised the max from 50 to 500, with the storage warning in docs/customisation.md (the top of the range is ~2.5 GB on disk at 2-5 MB per snapshot). |
 | Hundreds of identical `Trying to write N to X didn't complete got M` warnings overnight | Before concluding control is broken, check whether the planner even had a window in that period: the writes may be the routine idle reset — `execute.py` re-writes charge/discharge rate to max every cycle while not charging so PV can still charge (GH#5073, where the planner's own overnight windows were empty and no charging was lost — pure log noise). `write_and_poll_value()` has no failure dedup: after `INVERTER_MAX_RETRY` attempts it warns once, flags `had_errors` and clears the ledger entry, so the next cycle re-attempts from scratch — a firmware state that permanently refuses a register (SolarEdge preserve-charge holding its charge-limit register at 0 is the first confirmed instance) produces ~10 refused writes + a warning per cycle, indefinitely; any brand with a refuse-silently register maps here. Exception since PR #5297: on the `GWMQTT` (Predbat hub) type the retry policy is 3 attempts, and a same-target control failing twice degrades to one write per 5 minutes until it verifies - the "indefinitely" reading is a non-gateway-type reading only. **A stable quantised read-back below the verify tolerance never converges (GH#5324, verified on main):** the entity-write tolerance is `battery_rate_max_charge * MINUTE_WATT / 20` — 5% of the **allocated** rate, and `battery_rate_max_charge = min(inverter_limit_charge, battery_rate_max_raw)/MINUTE_WATT` (`inverter.py`) — so an `inverter_limit_charge` cap (including the API override, which is in `CONFIG_API_OVERRIDE`) below ~5 register steps makes a quantised read-back (GE: a 1300 W write sits stably at 1206) fail verify forever: 10 retries plus a full repeat every cycle, while the identical write at the uncapped rate verifies (any allocation below ~1.9 kW on a register-quantised rate is the risky band). The REST layer's tolerance for the same write is sized differently — `write_tolerance_watts()/12` from `Invertor_Max_Bat_Rate` with a 2600 W fallback (`givtcp_rest.py`) — and is immune to the cap, so two layers with different tolerances answer one write and the narrower one decides. The signature is a read-back **stable** at the same value off-target every cycle (never self-heals) - distinct from the settling bounce above, where the value sits between old and target only transiently. Workaround keeps the cap exactly: pick a register-exact override value. **Fixed in PR #5348 (v9.3.4):** `rate_tolerances()` (`inverter.py`) now widens the read-back tolerance **below** the written rate to one hardware register step — `fuzzy_below = max(fuzzy, step + 1)` where step is `INVERTER_DEF`'s `rate_step_percent_of_capacity` (GE ships 1) × `nominal_capacity` — mirroring the REST layer's step, and the GivTCP REST and Gateway MQTT paths got the same rule (`49b90af6`, `e7ca9b6a`), so a rounded-down quantised read-back within one step of the write verifies instead of restarting the retry loop; a read-back over the rate is never that rounding and keeps the plain fuzzy (one step **up** is still written down to a 0W hold), and where no `rate_step_percent_of_capacity` is declared the pre-#5348 5%-of-allocated-rate fuzziness still decides. Two case notes from that investigation: the preserve-mode hypothesis could not be verified from the dump because Predbat reads no SE battery-state entity, so the preserve flag is invisible in debug dumps; and a second execute pass ~46s after the first roughly doubled the warning volume in that log — observed only, mechanism never pinned down. `'Connection to inverter ID N failed'` on SE comes from the HA SolarEdge Modbus Multi integration, not Predbat. |
+| Too many inverter register writes / replans while the car charges, but the plan looks right | Hour-bucket the actual writes first to split car-conditional churn from a version regression: `grep "write_and_poll_value: Wrote" <log>` (the "No write needed" lines are noise), sed out the hour buckets, `sort | uniq -c`, and date any A/B from the`version X currently running` banners + `Predbat: Startup predbat` restart lines (GH#5434, 2026-10-07: 2-5 writes/h without the car on *both* v9.3.5 and v9.3.6, 9-24/h during the car's IOG session — car-conditional, not version-conditional, despite a mid-log upgrade). Car-conditional mechanisms, all log-verified there: (1) **dynamic-load flap** — the charger throttling load across `threshold_battery` flips `dynamic_load()` (plan.py, the `Dynamic load status changed from ... to ...`→ return True) which forces a replan (`predbat.py`,`Dynamic load adjustment changed, will recompute the plan`); (2) **`watch_list` event-entity refires** — a watched `octopus_free_session` event entity re-reports on the integration's refresh cycle (~15 min) and `watch_event()` (userinterface.py) sets `plan_valid = False` on *any* report, an event entity has no comparable old/new state so every timer bump looks new — unscheduled full replans; (3) **flat-band limit churn** — inside a near-flat cheap band overnight charging is metric-nearly-flat, so the adopted limit swings with small input moves under `metric_min_improvement_plan` (2p default) and both `should_replace_plan()` branches, the same tie surface as GH#5330; (4) **low power mode doubles each limit flip** — with `set_charge_low_power` on, a new target re-runs `find_charge_rate()`'s low-power search, rewriting charge_rate and the timed charge current together; (5) **window wipes** —`disabled charge window while waiting for schedule` (execute.py) when the plan re-picks a window start further out than `set_window_minutes` while the inverter's own window had already started, flipping scheduled charge off then back. Ruled out here: #5403's surplus replan — grep `Octopus Intelligent: dispatch minutes no car needs changed to N` before reaching for it. The "Best charge window" print at the top of a recompute is the prefilled incumbent, not the optimised plan — see the attachment-forensics trap. |
 | Low-rate sensors stuck on / the day rate classed as low rate while a saving session or Axle event is in the forecast | The auto threshold in `set_rate_thresholds()` (fetch.py) is computed from boost-contaminated stats. Both event loaders write the reward into the *import* rate table too, tagging the minutes `saving` in `rate_replicate`, and `rate_scan()` runs afterwards — so with `rate_low_threshold: 0` the auto branch's `rate_max - 0.5` rides on a `rate_max` that includes the boosted minutes: a +100p event on a 25.95/3.49p tariff sets the import threshold to 125.45p and every window classifies as low (GH#5050, reproduced in a test through `rate_scan` → `set_rate_thresholds` → `rate_scan_window`). `rate_average` is inflated too (30.54 vs ~17.7 true in the repro), so the *manual* multiplier path is not a clean workaround either — it must be picked against the inflated `rate_average` the log's `Import rates: min/max/average` line prints, and in the repro even `0.9` still lands above the day rate (27.49 > 25.95). The in-cycle auto-correction (`rate_import_cost_threshold = highest` after `rate_scan_window`) fixes the threshold but comes after `low_rates` was already built with the wrong one, and every same-cycle consumer reads the list — the low-rate sensors (output.py), the car plan, and `charge_window_best = clone_windows(low_rates)` when `set_charge_window` is on; the price-level search re-prices from per-minute rates so plan outcomes are usually near-correct, it is the candidate granularity that degrades. The export side is mirrored: the auto export branch cross-compares `rate_export_min` against import `rate_min`, both boost-contaminated. Fix key: both loaders already tag the boosted minutes `saving` in `rate_replicate`, so derive the threshold stats from untagged minutes — no new tracking needed; the in-tree precedent for keeping event prices out of a stat is `rate_export_max_forward`, deliberately computed from `rate_export_base` before the boosts. **The import side has the same distortion through the scan overwrite, and it *widens* the cheap band (GH#5237, code-read + `car_charging_smart` suite 11/11 on main):** the stamp precedes the scan (`load_saving_slot()` runs before `rate_scan()`), so `rate_max` includes stamped minutes and the provisional `rate_max - 0.5` threshold rises with it — then the scan overwrite (`rate_import_cost_threshold = highest`, fetch.py) admits minutes that fail the provisional but pass the final: on the #2716 arithmetic a session pushed `rate_max` 25.2→34.6p and the plain 25.2p peak entered `low_rates`. `low_rates` feeds **both** car planning and the `charge_window_best` seed (`plan.py`), so battery charge-window candidates are affected too. Fix direction already in-tree: `rate_max_base` (pre-stamp max, `fetch.py`) exists, and the export side already uses the base-tariff concept for exactly this class (`rate_export_max_forward_calc`). PR #5163 (open) rewires the stats to event-free (`rate_minmax_excluding_saving`) but does not touch the overwrite itself. Also re-verified from the same issue: the high-price charge gates (`plan.py`) skip only while `charge_limit_best[window_n] == 0` — already-limited windows are exempt, so "the gate let a high-price window through" reports must first check whether an earlier pass gave it a limit. |
 | Energy totals jump or reset | Provider-side cumulative counters and any monotonic clamping. **Check the counter's read pattern for a dip-and-recovery first (GH#5388, 2026-10-04, code-verified):** `minute_data()`'s smoothing branch pads a dip forward at the low value or interpolates clamped to `last_state`, but only zeroes an **upward** recovery when its per-minute step exceeds `MAX_INCREMENT` = 1.2 kWh/min (`const.py`) — so a dip held long enough that its recovery ramp is gentler than 1.2 kWh/min (gap of at least R/1.2 minutes for a recovery of R kWh) is counted in full as that day's energy (GH#5388's 181.3 kWh was exactly a recovery size), while a single-read upward jump is zeroed instead; the asymmetry is the signature. Applies to any integration binding `pv_today`/`load_today`/`import_today`/`export_today` to a cumulative counter; on SolaX the same shape was behind GH#5356/#5388 (see the SolaX row) and PR #5389 added a SolaX-side `hold_counter_dip()` — the `minute_data` asymmetry itself is unfixed and explains the symptom on other integrations. |
 | Works in HA, broken in Docker/standalone | `ha.py` websocket and `userinterface.py` callback return values. |
@@ -199,7 +201,7 @@ Grep for the named symbol rather than trusting a line number.
 | Predbat can't self-update, "unable to fetch" or a 403 | The unauthenticated GitHub API quota shared across the egress IP, not a broken release (GH#4886). |
 | A downloaded debug yaml has empty step arrays | Expected, not corruption. The arrays are built during `calculate_plan()` and freed at the end of the same `update_pred()` cycle whenever `debug_enable` is off (`predbat.py:1329-1334`), so a UI download taken afterwards never sees them. Deterministic on every download, not a race in the 2-hour auto-disable window (GH#4853). |
 | Memory usage high / addon RSS creeps up | Measured on Linux against a live config (no Load ML, ~40 threads): the idle floor is ~75 MB of imports before any data (stdlib 23, aiohttp 12, numpy 13 even with ML off, requests 6, own modules 13), live app data is only ~25 MB (HA state cache 8, minute arrays 5, rates 3, config 3), and the rest was allocator residue. Two things made most of the difference and are now in the tree: `yaml.dump()` builds a node tree ~30x the text (~150 MB for a full debug dict) so `create_debug_yaml()` streams one top-level key at a time through `DebugYamlDumper`, whose shared anchor counter is what stops the pieces failing to load with "found duplicate anchor"; and glibc keeps freed pages plus one malloc arena per thread, so `update_pred()` calls `malloc_trim()` after `gc.collect()` and `initialize()` caps arenas with `limit_malloc_arenas()` (both no-ops on macOS/musl). Baseline 266 MB peak / 162 MB idle became 138 / 126. If someone is still well above that, ask whether Load ML is on (adds ~30 MB idle plus 16 OpenBLAS threads) and check `Threads` in `/proc/<pid>/status`; the next lever is lazy imports (numpy when ML is off, components on enable), not the state cache. |
-| Plan seems stuck / doesn't adopt a clearly better replan | `should_replace_plan()` (`plan.py:1103`) compares pre-clip snapshots; a retained incumbent keeps its stale snapshot without the expired-window pruning `charge_window_best` gets, so the improvement gap can read identical for hours while real costs move. Combined with the fragmentation tie-break (`plan.py:1114`), which always favours the plan that adds no slot, a genuinely better plan can stay rejected until something else invalidates the plan outright. Confirmed from a 24h log: 180 consecutive comparisons with an unchanged gap, only 7 of 187 total ever adopting a new plan (GH#4736). Not a bug in the tie-break itself — `test_plan_tiebreak.py` asserts exactly that rule on purpose. **Every one of these gates is absolute in the rate sensor's own unit, so the tariff's scale multiplies how hard they bind (GH#5261, verified by code read + parsing a plan-HTML attachment 2026-09-27):** `currency_symbols` is display/labelling only — the metric and every margin gate run in the raw unit of the rate sensor, so a tariff sensor scaled to c/kWh (the European templates that publish rates ×100, e.g. on Iberian dynamic tariffs) makes `metric_min_improvement_plan` (default 2.0, ≈ €0.02 there) sit against slot prices of 0.02-0.05 — the #4736 mechanism amplified ~20x, and `should_replace_plan()` never adopts a ~0.1-unit-better plan; the within-run `metric_min_improvement` defaults to 0.0, so retention is the binding gate. Start any cent-scale-or-negative-price "why didn't the plan X" report by computing the stake in the sensor's unit before comparing against a gate. A plan-HTML export attachment is parseable without a debug yaml: rows carry `data-minute`/`data-rate` on the import/export cells, merged charge windows appear as `rowspan` "Chrg" cells, and `rate_range_text()` renders the tooltip of a merged cell as a `{min}-{max}` range — a `(-0.05-0.05p/kWh)` tooltip is one window spanning differing rates, not one price. |
+| Plan seems stuck / doesn't adopt a clearly better replan | `should_replace_plan()` (plan.py) compares pre-clip snapshots; a retained incumbent keeps its stale snapshot without the expired-window pruning `charge_window_best` gets, so the improvement gap can read identical for hours while real costs move. Combined with the fragmentation branch inside `should_replace_plan()` (PR #4301), which on a near-tie favours the less fragmented plan and even adopts a cost-neutral less-fragmented plan outright, a genuinely-better-by-less-than-the-gate plan can stay rejected until something else invalidates the plan outright. Confirmed from a 24h log: 180 consecutive comparisons with an unchanged gap, only 7 of 187 total ever adopting a new plan (GH#4736). Not a bug in the tie-break itself — `test_plan_tiebreak.py` asserts exactly that rule on purpose. **Every one of these gates is absolute in the rate sensor's own unit, so the tariff's scale multiplies how hard they bind (GH#5261, verified by code read + parsing a plan-HTML attachment 2026-09-27):** `currency_symbols` is display/labelling only — the metric and every margin gate run in the raw unit of the rate sensor, so a tariff sensor scaled to c/kWh (the European templates that publish rates ×100, e.g. on Iberian dynamic tariffs) makes `metric_min_improvement_plan` (default 2.0, ≈ €0.02 there) sit against slot prices of 0.02-0.05 — the #4736 mechanism amplified ~20x, and `should_replace_plan()` never adopts a ~0.1-unit-better plan; the within-run `metric_min_improvement` defaults to 0.0, so retention is the binding gate. Start any cent-scale-or-negative-price "why didn't the plan X" report by computing the stake in the sensor's unit before comparing against a gate. A plan-HTML export attachment is parseable without a debug yaml: rows carry `data-minute`/`data-rate` on the import/export cells, merged charge windows appear as `rowspan` "Chrg" cells, and `rate_range_text()` renders the tooltip of a merged cell as a `{min}-{max}` range — a `(-0.05-0.05p/kWh)` tooltip is one window spanning differing rates, not one price. |
 | Operation mode flip-flops between autonomous / self-consumption (or TBC / Self-Powered) within one control cycle and the export never engages — Tesla (and any service-template-controlled) inverter | **Fixed in PR #5114 (merged 2026-09-19)** — keep the mechanism for pre-#5114 logs. The `Charging/Discharging off via service` block (`execute.py`) had no `isExporting` guard: while an export is active `isCharging` is False, so it fired `adjust_charge_immediate(0)` → `charge_stop_service` 1-4s after `adjust_export_immediate()` had already run charge_stop + discharge_start, and on the shipped Tesla template both write the same operation-mode select, so the trailing write reverted the mode every cycle (GH#4165 and GH#4641, verified from both reporters' logs). #5114 skips the block while exporting — `adjust_export_immediate()` already stops charging as part of starting the export — and the `run_execute_test` status assertions now assert the charge-off never fires on `Exporting`/`Freeze exporting` statuses. Pre-fix log signature: `Calling service charge_stop_service` in a second pair *after* the discharge_start pair, within one 5-min cycle. On pre-fix versions `call_service_template`'s dedup would suppress the identical trailing call, but the shipped `templates/tesla_powerwall.yaml` sets `repeat: True` on every select write, which bypasses it — removing `repeat: True` from `charge_stop_service` is the user-side workaround. #4165 is the original; #4641 the dup. |
 | Charge and hold writes alternate every 5 minutes (GivTCP/GE "Setting Charge Schedule to enable/disable" + pause-mode toggles swapping nightly) while the plan target is rock-stable | The reported SoC is flipping by exactly 1% between cycles across a mid-quantisation-band boundary, not plan churn — in the verified case the target held at 71% for hours (a fractional limit, 6.64/9.39 kWh = 70.7%, printed "71%") while GE's SoC read 71↔70. `execute.py`'s hold branch gates on `soc_percent >= inv_target_soc_percent` with no tolerance; the intended 1% hysteresis (`soc_percent >= inv_target_soc_percent - 1.0`, comment "or within 1% below it") is nested *inside* that gate, so it is always true when reached and its else (re-enable the charge window) is dead — any SoC even 0.1% below target falls to "Charging" and re-enables the window, and next cycle at target it holds again (GH#5083, structure verified on main). Quick confirmation from their log: grep `Charge window will be` and check the target stays constant while current SoC alternates ±1% — read the raw lines, because the strings "current SoC X%," and "target Y%" shift field positions and an awk one-liner can make the target *look* like it flips. The gate/dead-hysteresis structure predates July's #4164 and the reporter's v8.52→v8.53.3 onset window touched neither the hold path nor the plan gate — late-Aug onset was situational (a target landing dead-heat on the quantisation band), not a regression. Fix direction: widen the gate to `target - 1.0`. Related, not duplicates: #4232 (EMS charge-target spam), #3693 (bulk GivTCP writes), and `soc-kw quantised by percent sensor` (same class, different sensor path). |
 | An inverter disappears from the plan after upgrading to v9.x (mixed fleet) | Start at the vendor component's `automatic_config()` and its `set_arg_auto(..., overwrite=True)` writes, not `execute.py`'s inverter loop. GH#5029: since the GivTCP REST refactor (#4864, v9.0.0+) GivTCP's auto-config wrote `inverter_type = ["GE" x discovered]` and `num_inverters = len(discovered)` (`givtcp.py`), so in a mixed fleet (GE on GivTCP REST + another brand configured manually) the manual `num_inverters` was shrunk and the non-GE inverter silently dropped out — the component-vs-manual instance of #4824. **Fixed in PR #5037 (v9.0.2)**: GivTCP's auto-config now only ever *raises* `num_inverters` (`_configured_num_inverters()` takes the largest value already configured) and `_keep_configured_tail()` preserved the configured list entries for any inverter beyond the discovered ones. PR #5216 (GH#5209) replaced `_keep_configured_tail()` with `_per_endpoint_values()`, which pins Predbat inverter n to REST endpoint n: the tail past the highest discovered endpoint is still kept as configured, but a gap slot below it (an endpoint that has not answered) keeps its configured entry or, with nothing configured, is bound to that endpoint's own not-yet-published entities, and `num_inverters` covers at least `max(discovered) + 1`. Keep the mechanism for pre-v9.0.2 logs: grep the reporter's full startup log for `auto-discovery wins` to confirm without needing their apps.yaml. `num_inverters`/`inverter_type` are **not** in `GIVTCP_AUTO_CONFIG_USER_WINS_KEYS` (that tuple is only the energy-today history keys, #4959) — don't cite that list as a guard. Fallback workaround at any version: `givtcp_automatic: False` (entities still published, auto-config skipped). Maintainer's preferred direction for the wider class: a per-component inverter offset, same spirit as the myenergi per-device-kind flag preference. |
@@ -249,7 +251,7 @@ Grep for the named symbol rather than trusting a line number.
 - **Building a component through its real constructor in a test aliases `self.args` to the live shared args** (found reviewing PR #5211's Apps-tab filter box). `ComponentBase.__init__` sets `self.args = base.args` — the same object the running fixture uses — and then runs `initialize(**kwargs)`, so a component constructed with `__init__` in a test is bound to the live shared args until a later line rebinds it, and silently renders whatever the shared fixture happens to hold if that rebinding is dropped or reordered. `WebInterface.initialize()` additionally constructs `AnnualPage` and `WebChat`, none of which a single-page test needs. The established pattern (`_reset_fixture`, `tests/test_web_apps_edit.py`) is `WebInterface.__new__(WebInterface)` plus explicit `base`/`log`/`prefix`/`args` — `html_apps()` needs exactly those four (it assigns `self.default_page` itself on entry). Generalises to all 18 components; same family as the `Execute.__new__` stub trap above from the other side (the stub misses base methods, the real constructor brings too much with it). Whether any currently-registered test actually mis-renders because of this is unaudited — suspected, not verified.
 - **Check that a report's premise exists before verifying its logic.** Three reports in two weeks were built on configuration or history that is not there: GH#4918 quoted `tesla_tariff_sync`, `tesla_smart_plan` and `tesla_smart_plan_max_export`, none of which appear anywhere in the codebase or `APPS_SCHEMA` (an AI assistant invented them for the reporter); GH#4917 treated `trust_future_dynamic_iog_slots` as shipped when it exists only on the closed-unmerged PR #4885 - on `main`, `rate_add_io_slots()` trusts every planned IOG slot unconditionally (pre-#5403 reading; since PR #5403 the `octopus_intelligent_consider_full` car-need gate withholds car-needless *future* dispatch minutes, though the current half hour still stamps — see the GH#5424 row); and GH#5022 (an enhancement issue written by Claude under the reporter's account) claims #5015 "turned off `support_feedin_first` for FoxCloud", which `git log -S` disproves — the flag was set once by `3ed9151e` and never removed. Grep `config.py` / `APPS_SCHEMA` for every setting a report names, and check `git log` or `gh pr view` before accepting an "added for #NNNN" claim — including the background section of an issue you are about to implement. The same check covers gating premises: GH#5237's issue body claimed the car-planning mitigations (`car_charging_plan_smart`, `car_charging_plan_max_price`, `best_soc_keep`) are "expert gated", false on both v9.0.3 and main — the two car settings carry `enable: num_cars` only and `best_soc_keep` has no gate at all — so a docs change premised on it would have been wasted; start any "setting hidden behind expert mode" claim with the item's `enable:` field in `config.py`. The same check runs in the other direction: an enhancement proposal can describe symbols that live only on the PR branch it was filed from — GH#5063 proposed hardening a log-redaction invalidation cache that existed only on the then-open PR #5053 and not on main (it has since merged, and PR #5171 has since moved the cache into the `LogRedaction` mixin (`log_secrets.py`) — so the cache it describes is real on current main), and overstated its own audit (production direct mutations of `self.args` are ~6 sites in 5 files — `chat_tools.py`, `create_missing_arg()` in `inverter.py`, `userinterface.py`, the one bulk `clear()+update()` in `web.py`, `web_chat.py` — not "roughly 15 files"; the deye/fox/teslemetry grep hits are `__main__` CLI blocks writing `mock_base`).
 - **Cross-check an attached debug yaml against the attached log before believing either (GH#5335).** Reporters sometimes attach someone else's `predbat_debug.yaml` — picked up from a community thread, and the issue's own opening comment may say so ("the debug shows sensor X which isn't my sensor"). At the yaml's capture minute, compare its state dump (`soc_kw`/`soc_max`) and its resolved entity serials (the "Regular expression argument" lines) against what the reporter's own log says at that timestamp: a different SoC, different Zappi/GE serials or Octopus account ids mean the yaml is foreign, only the log is then usable, and the reporter's own yaml must be re-asked for before config-dependent conclusions. The yaml's config can also flatly contradict the issue body (a direct zappi sensor named where the body says a regex) — the file wins over the prose, but only once it has been proven to be theirs.
-- **Attachment forensics: numbered logs, rolling snapshots and download-time dumps disagree (GH#5344).** Numbered AppDaemon logs (`predbat.1.log`, `predbat.2.log`) are download-time snapshots that overlap, not rotation order — one issue attached `.1.log` with the issue body (ending Oct 1 18:30) and `.2.log` in a follow-up comment (running Oct 1 00:00 → Oct 2 06:44), so the lower number was the *older* archive; check head+tail timestamps of every numbered log before building a timeline across them. The rolling debug-history tarball's newest snapshot is a ~3-hourly capture and can miss the triggering event entirely (there the newest still predated the ~18:00 IOG dispatch arrival the reporter described); the single `predbat_debug.yaml` is download-time state, not complaint-time state; and each snapshot is the state at *its* capture minute as well (see the GH#5335 bullet above). Within one log tick of the plan log, a repeated `Best charge window`/`Raw charge windows` pair is the retained plan re-logged plus the 5-minute recompute (`Recompute is saving previous plan...`, plan.py) — not two plans running in parallel, and the retained-first ordering means the *later* line is the newer plan, not the earlier one. And a reporter's incident **table** is built from HA-history samples, not log lines (GH#5419): the `record_status` lines (`... target X%-100.0%`) in their `predbat.log` are 5-minute precise — re-derive the SoC trajectory from those before fitting any threshold to the table's timestamps.
+- **Attachment forensics: numbered logs, rolling snapshots and download-time dumps disagree (GH#5344).** Numbered AppDaemon logs (`predbat.1.log`, `predbat.2.log`) are download-time snapshots that overlap, not rotation order — one issue attached `.1.log` with the issue body (ending Oct 1 18:30) and `.2.log` in a follow-up comment (running Oct 1 00:00 → Oct 2 06:44), so the lower number was the *older* archive; check head+tail timestamps of every numbered log before building a timeline across them. The rolling debug-history tarball's newest snapshot is a ~3-hourly capture and can miss the triggering event entirely (there the newest still predated the ~18:00 IOG dispatch arrival the reporter described); the single `predbat_debug.yaml` is download-time state, not complaint-time state; and each snapshot is the state at *its* capture minute as well (see the GH#5335 bullet above). Within one log tick of the plan log, a repeated `Best charge window`/`Raw charge windows` pair is the retained plan re-logged plus the 5-minute recompute (`Recompute is saving previous plan...`, plan.py) — not two plans running in parallel, and the retained-first ordering means the *later* line is the newer plan, not the earlier one. **And `Best charge window` is the prefilled incumbent, never the optimised plan** (GH#5434): the recompute block resets `charge_window_best` from `low_rates`/`charge_window` and runs `prefill_charge_limit_best()` before the line prints (plan.py), so it shows the incumbent limits carried onto the reset windows; the optimiser's output only appears in the later `Raw charge windows` / `Unclipped/Unfiltered/Filtered charge windows` lines, and the adoption decision is its own pair of lines after a metric comparison of both plans re-scored on the same cycle's inputs (`run_prediction_metric()` for each) — comparing the Best print against the writes that follow misreads which line is the plan.
 - **`grid_power` is negative when importing.** That is the repo's convention: `docs/faq.md` states it, `web.py`'s power-flow code reads it that way, and `test_enphase_api.py` asserts it by name - `grid_power_invert` exists precisely for meters that do the opposite. The outlier is `predbat_metrics.py:101`'s Prometheus help string, which claims "positive for import". GH#4797 read that string as authoritative and concluded the web arrows and icons were inverted; measured against every other statement of the convention, they are not. Work out which side of that contradiction a sign report is arguing from before agreeing with it.
 - **A passing test proves nothing when its fixture encodes the bug.** This has now happened often enough to be a check rather than a surprise: `web_power_flow` shares the sign assumption it would have to catch; the `iboost_smart` fixture uses a flat rate, so a rate-averaging bug cannot show; `test_balance_inverters` configures two *identical* inverters, so neither the capacity-blind SoC split nor the 3+ inverter partner ring is exercised (that was pre-#5149; the balance rework added `test_balance_pure.py`, which covers the SoC branches, cross-charge and pause semantics - and GH#4856 remains the canonical catalogue of balancer-design observations (ring partner selection, bang-bang control, equal-SoC% targets, the throttle-only surface): link to it rather than re-deriving them, and note pre-rework `balance_inverters()` line references are meaningless now that it is a pure function in `utils.py` driven from `rebalance_inverter_rates()`); PR #4935 found a cap test whose slot started after the window it was capping against, so the loop under test ran zero iterations; and GH#4826's `write_failure` test mocks a response with no error code, leaving the classification branch untested. `TestHAInterface.get_state` (test_infra.py) is its own instance of the same lesson: it returns `""` - not the caller's `default` - when a queried *attribute* is missing, while production `ha.py` returns the default, so a test written against the mock can be vacuous for default-handling when both values land in the same branch (found reviewing PR #5131). Read the fixture before quoting "the test passes" in a comment - and say what the test would have to look like to catch the reported behaviour. A sibling shape: **an allocator probe cannot test an identity-keyed cache guard** (GH#5171 review). The intuitive regression test for the `id()`-reuse bug in `LogRedaction._log_secret_fingerprint()` — free the mapping, allocate a batch of same-shaped dicts, assert none lands on the recorded address — **passed against the unfixed code** (64 decoy dicts after `gc.collect()`, no decoy ever landed there), because whether CPython recycles a freed dict's address depends on allocator pool state. What works deterministically is asserting the property that *rules the collision out*: the stored fingerprint still refers to the mapping it describes, so the recorded address belongs to a live object and cannot be recycled at all — `any(ref is stored for ref in gc.get_referrers(superseded_args))` (dicts are not weakref-able, so `gc.get_referrers()` is the tool); `test_log_secret_fingerprint_is_not_fooled_by_address_reuse()` in `tests/test_secrets.py` is the in-repo precedent, and the fix itself (hold the mappings, compare with `is`, not `==`) is merged (`7d17ea1b`). Generalises to any guard whose correctness rests on "this address still means what it meant": test that the object is pinned, not that the allocator declined to reuse the address.
 - **One unexplained observation, recorded so a second sighting is recognised.** In GH#4900's log, at 00:00:27, the nightly tariff-compare scenario for `igo_fixed` - running under swapped flat 12.0p export rates - had its scenario plan accepted as the **live** plan (`using new plan` logged from inside the scenario). A recalculation 18 seconds later replaced it, so nothing was harmed that night. The code path was not identified within triage scope. If you see `using new plan` inside a compare scenario, this is the thing.