Built-in Controllers Overview
What you'll learn: How Kubernetes uses specialized controllers to automate different aspects of cluster management.
What Are Built-in Controllers?
Remember from the previous lesson that the Controller Manager maintains your cluster's desired state? It does this by running multiple specialized controllers simultaneously—each responsible for a specific type of resource or task. Think of them like department managers in a company: each handles their domain, but all work toward the same organizational goal.
Common Built-in Controllers
Deployment Controller
Manages Deployments, ensuring the right number of pod replicas are running and orchestrating rolling updates. When you update an application, this controller gradually replaces old pods with new ones.
ReplicaSet Controller
Ensures a specified number of identical pod replicas are running at all times. If a pod crashes, this controller immediately creates a replacement. Deployments actually use ReplicaSets under the hood.
Node Controller
Monitors the health of worker nodes in your cluster. It detects when nodes become unreachable and triggers pod rescheduling to healthy nodes, typically after about 5 minutes of no communication.
Job Controller
Manages batch workloads that run to completion, like data processing tasks or database migrations.
DaemonSet Controller
Ensures specific pods run on every (or selected) node—useful for logging agents or monitoring tools that must exist cluster-wide.
How They Work Together
Each controller runs in a continuous loop: watch the current state via the API Server, compare it to desired state from etcd, and take corrective actions. They operate independently but coordinate through the shared API Server.
Key Takeaway: Built-in controllers are specialized managers within the Controller Manager that each maintain desired state for specific Kubernetes resources—like Deployment, ReplicaSet, and Node controllers—enabling automated, self-healing infrastructure.