Instead of a glossary, one small web app deployed to Kubernetes with five mistakes in it. Fixing each one in order explains what Pods, Deployments, ConfigMaps, Services and Ingress actually do, which kubectl command shows the problem, and how the pieces connect.
After we made a service run as a non-root user, it crashed on start with permission denied writing to its persistent volume. Adding fsGroup fixed that, then pods on a large volume took nine minutes to start. How volume ownership works in Kubernetes, fsGroupChangePolicy, and when an init container is the better tool.
We merged a hotfix minutes after a feature, and production ended up running the feature without the hotfix. Two pipeline runs deployed in the wrong order because the older one had a slower test stage. How we found it, and the fixes: concurrency groups, deploying immutable digests and refusing to deploy older commits.
A fix for an input validation bug was merged, confirmed in production, and two weeks later the bug was back. Nobody had reverted it. A long-running branch was merged with a conflict, and the resolution kept the old side. How we found the lost lines with git log, why the merge diff hid it and the habits that prevent it.