Example Documentation: Worker Queue
Goal
Trace asynchronous processing with a queue, producer, worker scaling and network boundaries.
What Is Done
The lab starts Redis as a queue, writes five tasks through a Job and processes them with multiple Worker Pods. An HPA describes target scaling, although actual metric evaluation requires Metrics Server.
Typical Commands
kubectl apply -f manifests/
kubectl get deploy,job,hpa,pdb -n queue-lab
kubectl logs -n queue-lab job/queue-producer
kubectl logs -n queue-lab deployment/queue-worker --tail=50
kubectl scale deployment queue-worker -n queue-lab --replicas=4
Validate Success
- Redis and Workers are ready.
- The Producer Job reaches
Complete. - Worker logs show processed
order-*entries. - NetworkPolicies allow Redis access only for Producer and Workers.
Common Mistakes
- Workers start before Redis and need several probe cycles.
- HPA shows
unknownwhen Metrics Server is missing. - Queue data is lost because Redis intentionally has no PVC in this lab.