kubernetes-patterns

Kubernetes deployment uses Deployment manifests to manage containerized workloads on a cluster. Define replicas, container images, and resource requests/limits in your manifest, then apply it with kubectl to orchestrate pod creation and scaling across nodes.

Liveness and readiness probes are health checks that monitor pod health. Liveness probes restart failed containers; readiness probes remove unhealthy pods from traffic. Both prevent serving requests from broken instances and improve reliability in production environments.

Services expose pods internally or externally using ClusterIP, NodePort, or LoadBalancer types. Ingress routes external HTTP/HTTPS traffic to Services based on hostnames and paths, providing a single entry point for managing external access to your cluster.

ConfigMaps store non-sensitive configuration data like environment variables; Secrets store sensitive data like credentials and tokens. Both inject configuration into pods at runtime without rebuilding images, enabling environment-specific deployments across development, staging, and production.

Use kubectl commands to inspect pod logs, describe resource states, and check events. Common troubleshooting targets image pull errors, resource exhaustion, health check failures, and scheduling issues—each requires examining pod status and events to diagnose root causes.

Horizontal Pod Autoscaling (HPA) scales replica counts based on CPU, memory, or custom metrics. Define target metrics and min/max replicas in an HPA resource; the controller automatically adjusts pod count to meet demand, optimizing resource utilization and cost.