Kubernetes Basic Concepts

#IaaS #Onboarding #Concepts

1. Basic Concepts

  • Kubernetes (K8s)

    A container orchestration tool. Automates container deployment, scaling, networking, and resource management.

  • Cluster

    The overall system in which Kubernetes operates.

    • Master Node (Control Plane): Responsible for cluster management (API server, scheduler, controller manager, etcd).
    • Worker Node (Node): Runs the actual containers (Pods).

2. Resource Units

  • Pod

    The smallest unit of container execution. Groups one or more containers together with storage and networking.

  • ReplicaSet

    A resource that maintains a specified number of Pods. (Guarantees n replicas)

  • Deployment

    Manages Pods + ReplicaSets in a declarative manner. Supports rolling updates and rollbacks.

  • StatefulSet

    For workloads that require state persistence (e.g., databases). Guarantees unique IDs and persistent storage.

  • DaemonSet

    A resource that runs one Pod per node (e.g., log collectors, monitoring agents).

  • Job / CronJob

    • Job: Executes a one-time task.
    • CronJob: Executes recurring tasks on a schedule.

3. Networking

  • Service

    An abstraction layer that provides a stable IP/domain to Pods.

    • ClusterIP (internal communication only),
    • NodePort (exposes a port on the node),
    • LoadBalancer (integrates with an external load balancer),
    • Headless Service (provides fixed DNS records).

  • Ingress

    Routes external HTTP/HTTPS requests to services inside the cluster. (Requires an Ingress Controller)

  • CNI (Container Network Interface)

    Pod networking plugins (Calico, Flannel, Cilium, etc.).

4. Storage

  • Volume

    Storage that can be shared between containers within a Pod.

  • PersistentVolume (PV)

    Storage managed at the cluster level.

  • PersistentVolumeClaim (PVC)

    A storage request made by a Pod.

  • StorageClass

    A storage policy for dynamic provisioning (e.g., AWS EBS, Ceph RBD).

5. Configuration & Security

  • ConfigMap

    Stores and injects non-sensitive configuration data.

  • Secret

    Stores sensitive data such as passwords and certificates. Uses base64 encoding.

  • Namespace

    Logically isolates resources.

  • RBAC (Role-Based Access Control)

    Controls permissions for users and service accounts.

6. Scheduling & Management

  • Scheduler

    Places Pods on appropriate nodes. Considers resource usage, taints/tolerations, and affinity.

  • Affinity / Anti-Affinity

    Pod placement policies (run alongside or away from specific nodes/Pods).

  • Taints & Tolerations

    Restricts scheduling so that only specific Pods can be placed on certain nodes.

  • Resource Requests & Limits

    Reserves and caps CPU/memory resources.

  • Horizontal Pod Autoscaler (HPA)

    Automatically adjusts the number of Pods based on CPU and memory usage.

  • Vertical Pod Autoscaler (VPA)

    Automatically adjusts Pod resource sizes (CPU/memory).

  • Cluster Autoscaler

    Automatically scales the number of cluster nodes up or down.

7. Logging & Monitoring

  • kubectl logs / kubectl exec

    Pod log viewing and debugging.

  • Prometheus / Grafana

    Metrics collection and monitoring.

  • ELK / EFK Stack

    Elasticsearch + (Fluentd/Fluent Bit) + Kibana logging.

8. Advanced Concepts

  • Operator

    A controller built to automate the operation of a specific application in the Kubernetes way.

  • CSI (Container Storage Interface)

    A standard storage interface. (AWS EBS CSI, Ceph CSI, PowerStore CSI, etc.)

  • CRD (Custom Resource Definition)

    Allows users to extend Kubernetes resource types.

  • Service Mesh (Istio, Linkerd)

    Provides inter-microservice communication, traffic management, security, and monitoring.