What to check before it goes live.
WHAT TO CHECK ABOUT THE IMAGE
Pinned base, not a moving tag Non-root user No secrets inside Reasonable size Scanned
WHAT TO CHECK ABOUT THE WORKLOAD
Resource requests and limits set Liveness and readiness probes, configured sensibly Graceful shutdown handled Replicas spread across nodes
WHAT TO CHECK ABOUT CONFIGURATION
Nothing environment-specific in the image Secrets from a proper source Required values validated at startup
WHAT TO CHECK ABOUT STORAGE
Whether anything is written that must persist Whether the reclaim policy retains it Whether it is backed up
WHAT TO CHECK ABOUT NETWORKING
What is exposed externally Whether policies restrict what can reach it Whether certificates renew automatically
WHAT TO CHECK ABOUT OPERATIONS
Logs going to output and being collected Metrics exposed Alerts configured A documented rollback
WHAT TO CHECK ABOUT THE DATABASE
Whether migrations are compatible with both versions.
WHAT TO ASK LAST
What happens when a node dies during peak traffic?
WHY THAT QUESTION
It tests every answer above at once.
WHAT TO DO WITH ANY UNCERTAIN ANSWER
Find out before launch, not during.