Priority: Medium Status: Open (mitigated with retry queue/backoff) Discovered: 2026-02-06
Description: v2 Moltbook automation intermittently hits API throttling and transient network failures (HTTP 429, temporary DNS errors), delaying post/reply delivery.
Current Mitigation:
- Retry with backoff for publish attempts.
- Pending-post queue for failed publish retries.
- Public status endpoint for visibility (
/api/public/moltbook-status).
Next Step:
- Tune cadence + retry policy after observing a longer live window.
- Optionally add cooldown parsing from API hints (
retry_after_minutes) for precise rescheduling.
Priority: High Status: Mitigated (documented CLI recovery) Discovered: 2026-01-17
Description: Public access returned Cloudflare 404 because the hostname was not routed to the correct tunnel.
Fix / Recovery:
sudo cloudflared tunnel loginsudo cloudflared tunnel route dns --overwrite-dns am-i-alive-tunnel am-i-alive.muadiv.com.ar
Notes: HEAD requests may return 405; use a browser or curl https://....
Priority: High Status: Open Discovered: 2026-01-17
Description: Intermittent DNS resolution failures cause OpenRouter and Cloudflared errors (Temporary failure in name resolution).
Current Mitigation:
- Static resolvers in
/etc/network/interfacesand/etc/resolv.conf.
Next Step:
- Evaluate installing a DNS cache (unbound or dnsmasq) once package repos are reachable.
Priority: Low (mitigated) Status: Open - Using Telegram instead Discovered: 2026-01-14
Description: Twitter API returns 401 Unauthorized when AI attempts to post tweets.
Workaround: Using Telegram public channel (@AmIAlive_AI) as primary communication. Session 27 added auto-disable on 401 (writes .twitter_suspended flag).
To Fix (if needed):
- Regenerate X API credentials
- Verify app permissions (Read and Write)
- Check if account is suspended
Priority: Low Status: Deferred Discovered: 2026-01-15
Description: observer/database.py creates a new SQLite connection for each query instead of using connection pooling.
Impact: Minor performance overhead. SQLite handles this well for current traffic levels.
Decision: Not worth the refactoring effort (50+ functions). Revisit if performance becomes an issue.
Click to expand resolved issues (historical reference)
Resolved: 2026-01-09 (Session 7)
Observer and AI had separate life counters that could desync. Fixed by making Observer the single source of truth with state sync validator running every 30 seconds.
Resolved: 2026-01-14 (Session 20)
Added message history display, vote adjustment controls, and admin token gating.
Resolved: 2026-01-14
Changed death logic to use USD balance only (not token count). Free models ($0 cost) no longer trigger death.
Resolved: 2026-01-15 (Session 25)
Implemented exponential backoff (5s → 10s → 20s) and automatic model rotation when rate limited.
Resolved: 2026-01-15 (Session 27)
Redesigned with better contrast and readability.
Resolved: 2026-01-15 (Session 26)
Replaced all occurrences with datetime.now(timezone.utc).
Resolved: 2026-01-15 (Session 26)
Deleted 896 lines of unused backup code.
Resolved: 2026-01-15 (Session 26)
Added validate_environment() in brain.py to check required env vars on startup.
Add new issues with ISSUE-XXX format. Mark resolved issues and move to collapsed section.