monitoring metrics_endpoint = ":9090" health_check_path = "/v1/health"
failover max_retries = 3 failover_timeout_ms = 30000 auto_recover = true acp hda node