Monitoring

Monitor queues, fetch failures, webhook failures, match proposal backlog, and observation freshness.

Suggested alerts:

  • Discovery jobs failing repeatedly.
  • Scrape latency or error rate above normal.
  • Webhook retry exhaustion.
  • No fresh observations for active high-priority targets.
  • Review queue growing faster than reviewers can clear it.
Audit trail

pi_fetch_logs and AI decision logs are operational data, not just compliance data.