{"title":"Agent Healthcheck & Liveness Probes","region":"Global","category":"Reliability","description":"Defining operational 'Vital Signs' to detect and restart hung or drifting agents.","lastUpdated":"2026-02-27","steps":["Implement a `/health` endpoint that checks for LLM API heartbeat and memory availability.","Set a `Liveness Probe` to detect 'Infinite Loop' hangs (no response for >60s).","Configure a `Readiness Probe` to ensure vector databases are indexed and reachable.","Monitor 'Token Velocity': Restart the agent if it exceeds 10k tokens in <10 seconds (loop detection).","Define a 'Graceful Shutdown' period to allow agents to finalize current tool-commits."],"url":"https://checklist.day/agent-healthcheck-liveness-probes"}