Production Debugging

Toggle DEBUG logging

Default is INFO. Send SIGUSR1 to the app process to flip between INFO and DEBUG at runtime. A WARNING log confirms each flip. Resets to the LOG_LEVEL env var on pod restart.

Kubernetes (PID 1 is the worker, --workers 1, no --reload):

# All backend replicas
kubectl get pods -l app.kubernetes.io/component=backend -o name \
  | xargs -I{} kubectl exec {} -- kill -USR1 1

Local dev (uvicorn --reload runs the worker in a multiprocessing child):

docker exec drone-tm-backend-1 pkill -USR1 -f multiprocessing-fork

Persist across restarts

Set env.LOG_LEVEL: "DEBUG" in your prod values file and sync (ArgoCD / helm upgrade). The chart renders it on both the backend and worker Deployments, which triggers a rolling restart.