Ingress Controller Architecture
What you'll learn: How Ingress Controllers actually implement HTTP routing rules by watching the Kubernetes API and configuring underlying proxies.
The Missing Piece
You've learned about Ingress resources (step 1219) — those YAML manifests that define routing rules like "send example.com/api to the API service." But those are just declarations. Something needs to actually read them and configure real networking infrastructure. That's the Ingress Controller's job.
How Ingress Controllers Work
An Ingress Controller is a Pod running in your cluster that:
- Watches the Kubernetes API for Ingress resources and Services
- Translates those resources into configuration for a reverse proxy (like NGINX, HAProxy, or Envoy)
- Updates the proxy's routing rules dynamically
- Handles incoming HTTP/HTTPS traffic based on those rules
Think of it like a restaurant manager watching orders (Ingress resources) and constantly updating the kitchen staff (proxy) about which dishes to prepare and where to deliver them.
The Two-Component Model
Unlike Services which are built into Kubernetes, Ingress needs two parts:
- Ingress Resource (YAML): What you want (routing rules)
- Ingress Controller (running Pod): What makes it happen (proxy configuration)
Different clusters can use different controllers (NGINX Ingress Controller, Traefik, HAProxy, cloud-specific ones), each with its own proxy implementation and features.
Why This Design?
The Ingress resource provides a standard interface, while controllers compete on implementation quality, performance, and features. You write the same Ingress YAML regardless of which controller you choose.
Key Takeaway: Ingress Controllers are active components that continuously watch Ingress resources and translate them into real proxy configurations, bridging the gap between declarative routing rules and actual HTTP traffic handling.