uvicorn workers were deadlocking invisibly both on startup and during GC. surviving workers within a pod still answered health checks so the pod just hung out indefinitely, degraded. fixed the metrics race, deleted a bunch of code trying to work around it and shipped one worker per pod