Skip to main content

Operations Runbook

This page is the published runbook for deployments, incidents, and monitoring. For production Supabase migration reliability, use the dedicated runbook:

CI/CD Checklist

  • Run lint, docs lint, typecheck, tests, and builds before merge.
  • Treat Mintlify link checks and docs build validation as part of the normal verification path.
  • Use release tags only after the repo contract is green.
  • Track container builds, migration validation, dependency review, and secret scanning as hardening backlog until they are added to the default workflows.

Common Incidents

  1. Webhook outage – inspect /internal/runtime, check verification errors, replay affected envelopes.
  2. Rate-limit breach – pause campaigns, adjust scheduler backoff, document 130429/131056 errors.
  3. Journey/campaign failure – review journey_runs, replay via /internal/webhooks/:id/replay, log in adversarial notes.

Monitoring Targets

  • webhook latency
  • queue depth
  • template validation failures
  • per-phone-number throughput
  • pair-rate backoff events
  • delivery error distributions
  • feature-flag drift

Routine Tasks

  • Rotate credentials, log audit events.
  • Update adversarial notes after fixes.
  • Sweep retention policies for raw payloads and rejections.

gb-agent connector triage

Use this block when an operator reports a stuck or surprising gb-agent draft state. Full surface details in Platform → gb-agent Connector.