Automated probes that verify an AI agent is running, responsive, and functioning correctly at regular intervals.
Health checks are the heartbeat monitor for your AI agent. They verify that the agent container is running, the process is responsive, the gateway is accepting connections, and the agent can reach its LLM provider.
Clawctl runs health checks at multiple levels: container health (is Docker running?), process health (is the OpenClaw process alive?), gateway health (is the API responding?), and functional health (can the agent process a test message?).
Failed health checks trigger the auto-recovery pipeline: restart, redeploy, alert. This ensures issues are caught and fixed automatically before they affect users.
An unhealthy agent that appears "running" is worse than a dead one. Users experience errors or timeouts without understanding why. Health checks catch these silent failures.
Clawctl runs multi-level health checks every 5 minutes. Dashboard shows real-time agent health status. Failed checks trigger automatic recovery. Historical health data helps identify recurring issues.
Try Clawctl — 60 Second DeployContainer status, process health, gateway responsiveness, LLM provider connectivity, and basic functional verification.
Every 5 minutes by default. Post-deploy checks run immediately with a 15-second warm-up period.
Yes. The Clawctl dashboard shows health status over time, helping identify intermittent issues.
Agent Recovery
Automated detection and correction of agent failures — including container crashes, health check failures, and degraded performance.
Agent Monitoring
Real-time observation of AI agent behavior, performance, and health — including conversation quality, error rates, and resource usage.
Production Readiness
The state where an AI agent meets all security, reliability, and operational requirements for serving real users with real data.
Agent Deployment
The process of provisioning infrastructure, configuring security, and launching an AI agent into a production environment.