Why HTTP Health Checks Can Silently Fail Background Queue Workers

A technical analysis highlights a critical blind spot in using standard HTTP /health endpoints to monitor background queue workers. Unlike request-handling services, where responding to a health check itself proves the service is working, a worker's HTTP listener operates independently of its actual processing loop. This means the endpoint can keep returning a healthy 200 status even while the worker's job queue is completely stalled due to deadlocks or exhausted connection pools. The orchestrator continues to see a green signal while the backlog quietly grows, making the health check structurally incapable of detecting the very failure it was meant to catch. The author argues that workers require a different monitoring approach, such as a heartbeat file that the worker actively rewrites as it processes jobs, rather than a borrowed convention designed for request-based services.