Background worker health status not exposed as metrics #962

Labels: monitoring, reliability

Priority: High

Description

Background workers (email queue worker, blockchain sync worker, rate limit cleanup, newsletter cleanup) run as tokio tasks with no health metrics. Operators cannot determine from Prometheus whether any worker is running, stuck, or has crashed.

Acceptance Criteria

• Add a Prometheus gauge worker_status{name} (1 = running, 0 = stopped) for each background worker
• Update the gauge in each worker loop heartbeat and clear it on panic/exit
• Add a Prometheus alert firing when any worker_status drops to 0 for more than 60 seconds