Automation rule engine
Match a trigger. Run a governed action. Every firing on the audit chain.
Per-project rules watch for a task or scheduler trigger and run a governed action: notify an operator, retry a task, or cancel a task. Every rule carries its own guardrails. A rolling-window circuit breaker trips a runaway rule into notify-only before it can amplify an incident, a per-project kill switch stops all automation at once, and a dry-run mode records exactly what a rule would have done without touching anything. Creating, editing, re-arming, or re-enabling destructive rules requires admin authority. Browser-session mutations also require fresh MFA; bearer-authenticated callers follow their API-key authority. At fire time the runner rechecks that the creator still has project-admin membership, and every firing lands on the HMAC-chained audit log.
Ships with
- Triggers on task events (failed, succeeded, retried, orphaned) and infrastructure events (worker offline, schedule misfired)
- Governed actions: notify an operator, retry a task, or cancel a task
- Rolling-window circuit breaker trips a runaway rule into notify-only automatically
- Per-project kill switch halts every rule at once; reset-circuit and delete from the dashboard
- Dry-run mode records what would fire without acting, so a new rule proves itself first
- A rule can condition on an issue fingerprint, so it fires for one specific failure signature
Highlights
- Destructive rules require admin authority; browser-session writes require fresh MFA, and fire time rechecks the creator's current admin membership
- Every firing writes to the HMAC-chained audit log (which rule, which trigger, what it did)
- Dashboard Automation area: rule list, create/edit builder, kill switch, reset-circuit, delete
- Notify-only actions stay available even when the breaker has tripped, so you never go dark