Kubernetes - Pods, Deployments & Services

The management layer of a Kubernetes cluster that maintains cluster state and coordinates worker nodes. Key components include:

A worker machine (physical or virtual) that runs containerized application workloads via Pods. Key components include:

The smallest, most basic deployable object in Kubernetes. A Pod encapsulates one or more co-located containers that share storage resources (volumes), a unique network IP, and container runtime settings.

A design pattern where a secondary helper container runs alongside the main application container inside the same Pod. Both containers share localhost networking and shared volumes, making it ideal for tasks like log collection, monitoring agents, or service mesh proxies (e.g., Envoy).

A Kubernetes controller object whose primary purpose is to maintain a stable set of identical replica Pods running at any given time. It uses a set selector to monitor Pod labels and creates or deletes Pods as needed to match the desired count.

A higher-level Kubernetes API object that declaratively manages ReplicaSets and Pods. It provides declarative updates, automated rollouts, scaling capabilities, and effortless rollbacks.

Deployments support two primary update strategies:
1. RollingUpdate (Default): Gradually replaces old Pods with new ones, ensuring zero downtime by controlling maxUnavailable and maxSurge instances.
2. Recreate: Terminates all existing Pods simultaneously before creating new ones, causing temporary downtime but avoiding dual-version execution.

The ability to revert a Deployment's revisions to a previous stable state using kubectl rollout undo deployment/<deployment-name>. Kubernetes tracks revision history by maintaining older ReplicaSets with replicas: 0.