Troubleshooting Resource Quotas and LimitRanges

A Deployment scales to 10 replicas, but only 7 pods appear. No crash loops. No OOM kills. kubectl get deployment shows 7/10 under READY and nothing else. The problem is invisible until you check the ReplicaSet events:

The rejection happens at admission, and the error surfaces only in events on the owning ReplicaSet or Job. Not on the Deployment. Not on the CronJob. The cka-resource-management lesson covers how these policies work. This lesson is about what happens when they break: reading the error messages the admission plugins produce, tracing the source of each rejection, and fixing the five most common failure patterns.