One active governor owns a ledger at a time, so throughput is a single process; --standby is failover, not scale-out.
Goal: let a fleet of governors share one logical budget - e.g. per-run sharding across processes, or a coordination layer over the ledger - so a shared team gateway scales past one process.
Constraints: must preserve the at-most-once counting and stopped-stays-stopped guarantees; Postgres already gives a cross-process lease to build on.
Tracked in docs/known-issues.md as a deferred item from the v0.2 architecture-review remediation.
One active governor owns a ledger at a time, so throughput is a single process;
--standbyis failover, not scale-out.Goal: let a fleet of governors share one logical budget - e.g. per-run sharding across processes, or a coordination layer over the ledger - so a shared team gateway scales past one process.
Constraints: must preserve the at-most-once counting and stopped-stays-stopped guarantees; Postgres already gives a cross-process lease to build on.
Tracked in
docs/known-issues.mdas a deferred item from the v0.2 architecture-review remediation.