Production Best Practices

Before You Go Live

• Set resource requests and limits on every container.
• Add liveness and readiness probes so Kubernetes restarts or reroutes broken pods.
• Encrypt Secrets at rest and rotate them regularly.
• Back up etcd and any persistent data; test the restore.

Probes Example

containers:
- name: app
  image: myapp:1.2
  livenessProbe:
    httpGet:
      path: /healthz
      port: 8080
  readinessProbe:
    httpGet:
      path: /readyz
      port: 8080

readiness gates traffic while a pod is still starting; liveness restarts a pod that hangs forever.

Declarative Everything

Keep every manifest in Git. Apply with kubectl apply or better, a GitOps tool (Argo CD, Flux) that syncs the cluster to your repo.

Versioning and Upgrades

• Pin container image tags; never rely on latest.
• Use backups and practice restores for etcd and databases.
• Upgrade components (Kubernetes, cluster add-ons, Helm charts) gradually and verify each step.

Security Essentials

• Run containers as non-root with read-only filesystems where possible.
• Use NetworkPolicies to restrict pod-to-pod traffic.
• Regularly run vulnerability scans on your images.

Key Points

• Probes + resource limits make workloads self-healing and safe.
• GitOps keeps the cluster reproducible and auditable.
• Security is a loop: images, access, network policies and secrets all need attention.