4 Commits

Include renames

Author	SHA1	Message	Date
Algis DumbrisandClaude Opus 4.6	7848911a5f	fix: ignore system DMs in reactor — prevent stalemate notification cascade ... The StaleWorker sends DMs from 'system' to all channel members when workflow messages are stuck in 'proposed' state. These DMs were triggering reactive agent runs, which couldn't action the stale messages, burning daily budget on wasted K8s Jobs. Now: reactor silently ignores all messages from 'system' sender. System notifications are for human review, not agent action. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-03-26 08:56:37 +02:00
Algis DumbrisandClaude Opus 4.6	64c68c22be	fix(014): prevent stuck runs by creating K8s Job before DB insert ... The reactor was inserting the run record first, then creating the K8s Job, then updating the record with the job name. If the update failed (SQLITE_BUSY), the run would be stuck in 'running' with no job name, making it invisible to the poller. Now: create K8s Job first, then insert the run record with job name already set in a single atomic write. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-03-26 06:09:16 +02:00
Algis DumbrisandClaude Opus 4.6	cf6066229f	feat(014): add Prometheus metrics, Grafana dashboard, volume mounts, resource tuning ... - Reactor Prometheus metrics: triggers_total, run_duration_seconds, agent_running, budget_used_today - Integrated promauto metrics into hand-rolled WritePrometheus endpoint - K8s runner: ImagePullPolicy=IfNotPresent, volume mounts, CLI args support - Reactor: 2Gi/500m default resources (agent SDK needs it), 1h timeout - Grafana dashboard "SynapBus Reactive Agents" with 8 panels: triggers by status, agent state, budget gauge, run duration, agent turns from Loki, reactor events log, agent container logs - SQLite: busy_timeout=15s, synchronous=NORMAL, MaxOpenConns=4 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-03-25 22:15:54 +02:00
Algis DumbrisandClaude Opus 4.6	6afe1853ad	feat(014): implement reactive agent triggering engine ... - Migration 015: extends agents with trigger config, adds reactive_runs table - Reactor engine: decision chain (mode, depth, budget, cooldown, sequential) - Reactor store: SQLite persistence for runs with RFC3339 timestamps - Reactor poller: K8s Job status polling (15s interval) - Failure notifier: system DM to owner on job failure - REST API: /api/runs, /api/runs/:id, /api/runs/:id/retry, /api/agents/reactive - Agent model: trigger_mode, cooldown, budget, depth, k8s_image, pending_work - K8s runner: GetClientset() for poller - All 28 test packages pass (8 new reactor tests) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-03-25 17:42:48 +02:00