Logs, Exec and Debug Containers

kubectl logs has several flags worth using deliberately rather than by habit: -c <container> picks a specific container when a Pod has more than one, --since=10m limits output to a recent window instead of scrolling through an entire history, and combining a label selector with logs lets you pull from every Pod in a set at once - though be aware that output from multiple Pods interleaves in the terminal without clearly separating which line came from which Pod. kubectl exec lets you open an interactive shell or run a one-off command inside an already-running container - but it depends entirely on that container's image actually containing a shell and useful tools, and many modern, security-hardened production images (deliberately, for the reasons covered in the security module) ship with neither. For exactly that situation, an ephemeral debug container lets you attach a separate, tool-rich container (like a general Linux troubleshooting image) into a running Pod's shared namespaces temporarily, without ever rebuilding or modifying the original application image at all. Debugging at the node level itself - rather than inside a specific Pod - is a distinct, higher-privilege escalation that should be treated with the same caution as any other elevated access.

Analogy: Kubernetes is a thermostat, not a remote control. You declare the temperature; independent controllers measure reality and keep acting. Debugging means finding which sensor, rule or actuator prevents convergence.

Read the object as evidence

kubectl get RESOURCE NAME -n NAMESPACE -o yaml
kubectl describe RESOURCE NAME -n NAMESPACE
kubectl get events -n NAMESPACE --sort-by=.metadata.creationTimestamp
kubectl explain RESOURCE.spec

Do not memorize these as a ritual. The first command exposes desired and observed state, the second connects conditions and events, the third supplies a timeline, and the fourth checks the server's schema. Compare what the controller was asked to do with what it reports doing. Then test the narrowest hypothesis.

Scenario: A Pod stuck in CrashLoopBackOff appears to have suspiciously empty logs whenever a responder checks kubectl logs, leading them to wrongly conclude the application is failing to log anything at all. What is actually happening is that kubectl logs by default shows the current container attempt, which for a rapidly crash-looping Pod may be only a fraction of a second old and never got the chance to log anything before dying again - the real evidence lives in the previous attempt. Adding --previous retrieves the logs from that last terminated attempt instead, which is almost always where the actual failure reason is recorded.

Choose the right observation path

Use kubectl logs POD -c CONTAINER --previous --since=10m for terminated attempts and label selectors for a replica set, while remembering logs from many Pods interleave. exec needs a running container and tools inside it. An ephemeral container can share namespaces for packet, process or filesystem inspection without changing the image, subject to security policy. kubectl debug node/NODE is a higher-privilege escalation.

Scenario: A distroless container has no shell, so exec fails with file not found. That is an image property, not proof the container is dead. Attach an approved debug image or observe externally.
Warning: Debug containers can expose process memory, network traffic and mounted secrets. Audit and remove their use like other privileged incident access.

Production reasoning

Ask four questions: Who owns this object? What dependency must become ready next? Which controller reports the blocking condition? What evidence would disprove my current theory? This prevents symptom-driven changes. Record the context, namespace, object generation, image digest and recent rollout before mutation; a recreated Pod may erase the evidence you needed.

Warning: Running is not the same as ready, healthy, durable or correct. Kubernetes status is layered. Confirm the application-level outcome as well as the object state.
Goal: Put this model into practice in the Kubernetes lab sidecar-redis-logs. Open /labs/kubernetes, choose sidecar-redis-logs, predict the failure path before changing anything, then use the simulator's check command to validate the finished state.

Deliberate practice

Before the lab, write the expected object relationship and the first three commands you will run. Afterward, explain why the fix converged and name one tempting change that would only mask the symptom. Repeat using an explicit namespace and a structured output format. This prediction-observation-explanation loop is what turns command familiarity into production judgment.

45-minute investigation

  1. Map (5 min): draw the owner-to-child chain and mark every namespace, selector, identity and dependency involved.
  2. Predict (5 min): write one expected status condition, one likely event and one log or metric signal before opening the lab.
  3. Observe (10 min): collect YAML, describe output and ordered events. Do not mutate state. Record which observation disproves your first theory.
  4. Repair (15 min): make the smallest declarative correction, watch the responsible controller converge, and verify the user-facing path rather than stopping at Running.
  5. Stress (10 min): change one relevant constraint - replica count, label, readiness, resource value or placement rule - predict the outcome, observe it, then restore the known-good declaration.
Tip: Keep a short incident note with symptom, evidence, hypothesis, change, result. Across four sections this produces a reusable runbook instead of a pile of remembered commands.