Service Lifecycle — io.oap.agents.lifecycle

The lifecycle capability adds operational controls on a named service. The {id} parameter is the service identifier as declared in the discovery manifest (services[].id).

REST API

Method	Path	Description
POST	/services/{id}/pause	Pause a running service
POST	/services/{id}/resume	Resume a paused service
POST	/services/{id}/heartbeat	Signal that a service is still alive

All return 204 No Content on success.

Heartbeat

Agents should call POST /services/{id}/heartbeat periodically to confirm they are still running. This prevents stale entries accumulating in the registry — if no heartbeat is received within the server-defined window, the server should transition the service status to "error".

The heartbeat interval is server-defined and should be documented by each implementation. A reasonable default is 60 seconds; agents should call more frequently than the window (e.g. every 30 seconds when the window is 60 seconds).

If an agent process crashes without a clean DELETE /services/{id}, the heartbeat mechanism ensures the registry self-corrects over time rather than accumulating ghost entries indefinitely.

Schema

See lifecycle.json.