The object map
Settings that decide reliability
| Setting | Why it matters |
|---|---|
readinessProbe | No traffic until the app is ready. Prevents errors during deploys |
livenessProbe | Restarts stuck processes. Don't point it at dependencies, or a DB blip becomes a restart storm |
resources.requests / limits | Scheduling and noisy-neighbour protection. A too-low memory limit causes OOMKilled |
PodDisruptionBudget | Keeps N replicas up during node drains |
topologySpreadConstraints | Spread pods across AZs |
Graceful shutdown (preStop, terminationGracePeriodSeconds) | Finish in-flight requests on SIGTERM |
Helm in one paragraph
Helm packages Kubernetes YAML as templated charts. values.yaml holds defaults, and per-environment values files override them. The same chart deploys every environment, which is Twelve-Factor "config in env" at cluster level.
Cheatsheets
The whole topic on one page. Click to open full screen.
Sources & further learning
Videos, courses, docs and books I recommend for this topic.
Related topics
CI/CD & GitOps
From commit to Kubernetes — build once, promote the same artefact, and let Git be the source of truth for what's running (ArgoCD).
Sidecar & Service Mesh
Move networking concerns — mTLS, retries, traffic splitting, telemetry — out of application code into a proxy deployed next to every service.
The Twelve-Factor App
Twelve rules for building cloud-native services that are portable, scalable and deployable anywhere — config in env, stateless processes, logs as streams.