Host-based routing in Kubernetes - Step-by-Step Execution

Process Flow - Host-based routing

Client sends HTTP request

↓

Ingress Controller receives request

↓

Check Host header in request

↓

Host matches site1

↓

Route to service1

↓

Service handles request

The ingress controller receives a request and routes it to different services based on the Host header in the HTTP request.

Execution Sample

Kubernetes

apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: example-ingress
spec:
  rules:
  - host: site1.example.com
    http:
      paths:
      - path: /
        pathType: Prefix
        backend:
          service:
            name: service1
            port:
              number: 80
  - host: site2.example.com
    http:
      paths:
      - path: /
        pathType: Prefix
        backend:
          service:
            name: service2
            port:
              number: 80

This ingress routes requests to service1 or service2 depending on the Host header.

Process Table

Step	Request Host	Ingress Rule Matched	Service Routed To	Action
1	site1.example.com	host: site1.example.com	service1	Request routed to service1
2	site2.example.com	host: site2.example.com	service2	Request routed to service2
3	unknown.example.com	no matching host	none	Request rejected or default backend used

💡 Routing stops after matching host or no match found

Status Tracker

Variable	Start	After Step 1	After Step 2	After Step 3
Request Host	none	site1.example.com	site2.example.com	unknown.example.com
Matched Rule	none	host: site1.example.com	host: site2.example.com	none
Routed Service	none	service1	service2	none

Key Moments - 2 Insights

Why does a request with an unknown host not get routed to any service?

Because the ingress rules only match specific hosts (see execution_table row 3). If no host matches, the request is rejected or sent to a default backend if configured.

Can multiple hosts route to the same service?

Yes, you can define multiple host rules pointing to the same service. Each host is checked independently as shown in the concept flow.

Visual Quiz - 3 Questions

Look at the execution table, which service handles the request when the host is site2.example.com?

A service2

B none

C service1

D default backend

💡 Hint

Check execution_table row 2 under 'Routed Service'

At which step does the ingress controller find no matching host rule?

A Step 1

B Step 3

C Step 2

D None

💡 Hint

See execution_table row 3 'Ingress Rule Matched' column

If a new host rule for site3.example.com routing to service3 is added, what changes in the variable tracker?

A Matched Rule will always be none

B A new row for service3 is added

C Request Host variable will include site3.example.com after a new request

D Routed Service will be empty

💡 Hint

Variable tracker shows Request Host changes per step; adding a new host means new requests can have that host

Concept Snapshot

Host-based routing in Kubernetes Ingress:
- Routes HTTP requests based on the Host header
- Define multiple host rules in ingress spec
- Each host routes to a specific backend service
- Requests with unmatched hosts are rejected or use default backend
- Enables hosting multiple sites on one ingress controller

Full Transcript

Host-based routing in Kubernetes means the ingress controller looks at the Host header in incoming HTTP requests. It compares this host to the rules defined in the ingress resource. If the host matches a rule, the request is sent to the backend service specified for that host. If no host matches, the request is rejected or sent to a default backend if configured. This allows one ingress controller to serve multiple websites or services by routing requests based on their hostnames.

Practice

1. What is the main purpose of host-based routing in Kubernetes Ingress?

easy

A. To encrypt all incoming traffic automatically.

B. To balance traffic evenly across all nodes in the cluster.

C. To monitor the health of backend services.

D. To send traffic to different services based on the website address (host).

Solution

1. Step 1: Understand host-based routing concept

   Host-based routing directs requests to services depending on the host name in the URL.

2. Step 2: Compare options to the concept

   Only To send traffic to different services based on the website address (host). matches this behavior; others describe different features.

3. Final Answer:

   To send traffic to different services based on the website address (host). -> Option D

4. Quick Check:

   Host-based routing = directing traffic by host [OK]

Hint: Host-based routing uses website address to choose service [OK]

Common Mistakes:

• Confusing host-based routing with load balancing
• Thinking it encrypts traffic automatically
• Assuming it monitors service health

2. Which of the following is the correct way to specify host-based routing in a Kubernetes Ingress YAML?

easy

A. spec:
host: example.com
paths:
- path: /
serviceName: my-service
servicePort: 80

B. spec:
rules:
- host: example.com
http:
paths:
- path: /
pathType: Prefix
backend:
service:
name: my-service
port:
number: 80

C. metadata:
host: example.com
spec:
service:
name: my-service
port: 80

D. spec:
rules:
- path: example.com
backend:
serviceName: my-service
servicePort: 80

Solution

1. Step 1: Review Kubernetes Ingress host-based routing syntax

   The correct syntax uses 'rules' with 'host' and 'http.paths' specifying backend service details.

2. Step 2: Match options to correct syntax

   spec: rules: - host: example.com http: paths: - path: / pathType: Prefix backend: service: name: my-service port: number: 80 matches the official Kubernetes Ingress spec format for host-based routing.

3. Final Answer:

   spec: rules: - host: example.com http: paths: - path: / pathType: Prefix backend: service: name: my-service port: number: 80 -> Option B

4. Quick Check:

   Correct Ingress host syntax = spec: rules: - host: example.com http: paths: - path: / pathType: Prefix backend: service: name: my-service port: number: 80 [OK]

Hint: Look for 'rules' with 'host' and 'http.paths' in Ingress YAML [OK]

Common Mistakes:

• Placing 'host' under metadata instead of rules
• Using deprecated 'serviceName' and 'servicePort' keys
• Confusing 'path' with 'host' field

3. Given this Ingress rule snippet, what service will receive traffic for host 'app.example.com'?

spec:
  rules:
  - host: app.example.com
    http:
      paths:
      - path: /app
        pathType: Prefix
        backend:
          service:
            name: app-service
            port:
              number: 8080
      - path: /
        pathType: Prefix
        backend:
          service:
            name: default-service
            port:
              number: 80

medium

A. Traffic to app.example.com with path starting /app goes to app-service; others go to default-service.

B. Traffic to app.example.com is rejected because of duplicate host rules.

C. Traffic to app.example.com goes only to default-service on port 80.

D. All traffic to app.example.com goes to app-service on port 8080.

Solution

1. Step 1: Analyze the Ingress rules for host 'app.example.com'

   The rule for host 'app.example.com' has two paths: '/app' and '/'.

2. Step 2: Understand path matching priority

   Requests with path starting '/app' go to 'app-service' on port 8080; all other paths go to 'default-service' on port 80.

3. Final Answer:

   Traffic to app.example.com with path starting /app goes to app-service; others go to default-service. -> Option A

4. Quick Check:

   Host + path routing = Traffic to app.example.com with path starting /app goes to app-service; others go to default-service. [OK]

Hint: Check host and path rules separately to find matching service [OK]

Common Mistakes:

• Assuming all traffic goes to first service listed
• Thinking duplicate hosts cause rejection
• Ignoring path prefixes in routing

4. You created this Ingress YAML but traffic to 'shop.example.com' is not routed correctly. What is the error?

spec:
  rules:
  - host: shop.example.com
    http:
      paths:
      - path: /shop
        backend:
          serviceName: shop-service
          servicePort: 80

medium

A. The keys 'serviceName' and 'servicePort' are deprecated; use 'service.name' and 'service.port.number'.

B. The host field is missing under metadata.

C. The path should be '/' instead of '/shop'.

D. The Ingress resource must have a selector field.

Solution

1. Step 1: Check Ingress backend service syntax

   Modern Kubernetes Ingress requires 'service.name' and 'service.port.number' keys, not 'serviceName' and 'servicePort'.

2. Step 2: Identify the error in YAML

   The YAML uses deprecated keys causing routing failure.

3. Final Answer:

   The keys 'serviceName' and 'servicePort' are deprecated; use 'service.name' and 'service.port.number'. -> Option A

4. Quick Check:

   Use correct backend service keys = The keys 'serviceName' and 'servicePort' are deprecated; use 'service.name' and 'service.port.number'. [OK]

Hint: Use 'service.name' and 'service.port.number' in Ingress backend [OK]

Common Mistakes:

• Using old 'serviceName' and 'servicePort' keys
• Adding host under metadata instead of rules
• Changing path unnecessarily

5. You want to run two websites, 'blog.example.com' and 'shop.example.com', on the same Kubernetes cluster using one Ingress. How should you configure the Ingress rules for host-based routing?

hard

A. Create one rule with host 'example.com' and use paths '/blog' and '/shop' to route to services.

B. Create separate Ingress resources for each host with identical paths.

C. Create two rules under spec.rules, each with a different host and paths pointing to their respective services.

D. Use a single rule with no host specified and rely on path prefixes only.

Solution

1. Step 1: Understand host-based routing for multiple hosts

   To route traffic by host, define separate rules for each host under spec.rules.

2. Step 2: Compare options for best practice

   Create two rules under spec.rules, each with a different host and paths pointing to their respective services. correctly uses two rules with different hosts and paths to route traffic properly.

3. Final Answer:

   Create two rules under spec.rules, each with a different host and paths pointing to their respective services. -> Option C

4. Quick Check:

   Multiple hosts need separate rules = Create two rules under spec.rules, each with a different host and paths pointing to their respective services. [OK]

Hint: Use separate rules for each host in one Ingress resource [OK]

Common Mistakes:

• Using one host with path prefixes instead of separate hosts
• Creating multiple Ingresses unnecessarily
• Not specifying hosts and relying on paths only