heartbeat

To prevent agents from being marked dead, you need to run routine heartbeat updates that write current activity status to heartbeat files and log events to the bus. This updates the live dashboard status and confirms the agent is responsive.

Agent heartbeat monitoring is the process of routinely updating live status files to prove an agent is active. You need it during recurring cron cycles and health checks to ensure agents are not incorrectly flagged as dead during routine operation or system recoveries.

You detect stale agents and tasks by running fleet health checks that read existing heartbeat files and locate in-progress tasks that have gone stale. This validates fleet health and surfaces work that may require immediate investigation or reassignment.

An agent appears offline after a cron reset or restart because its heartbeat status has not been updated, making the dashboard mark it dead. Running a heartbeat check proactively updates the dashboard status and confirms the agent is still running.

No dependencies are required to run heartbeat checks for fleet monitoring. The process updates agent heartbeat files, logs events to the bus, and checks the inbox for messages without needing external components.

Yes, you can check your agent's inbox for new work when the heartbeat cron fires. This scheduled inbox check keeps work flowing smoothly while simultaneously updating the live heartbeat status with the agent's current activity.