A/B testing with Ingress in Kubernetes - Time & Space Complexity

Time Complexity: A/B testing with Ingress

O(n)

Understanding Time Complexity

When using Kubernetes Ingress for A/B testing, it's important to understand how the routing rules affect processing time.

We want to know how the number of routing rules impacts the time it takes to handle incoming requests.

Scenario Under Consideration

Analyze the time complexity of the following Ingress configuration snippet.

apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: ab-testing-ingress
spec:
  rules:
  - host: example.com
    http:
      paths:
      - path: /v1
        pathType: Prefix
        backend:
          service:
            name: service-v1
            port:
              number: 80
      - path: /v2
        pathType: Prefix
        backend:
          service:
            name: service-v2
            port:
              number: 80

This Ingress routes traffic to two different service versions based on the URL path for A/B testing.

Identify Repeating Operations

Identify the loops, recursion, array traversals that repeat.

• Primary operation: Checking each path rule in the Ingress to find a match for the incoming request.
• How many times: The number of path rules (here 2) determines how many checks happen per request.

How Execution Grows With Input

As the number of path rules grows, the Ingress controller checks more rules to find the right backend.

Input Size (n)	Approx. Operations
10	10 path checks
100	100 path checks
1000	1000 path checks

Pattern observation: The number of checks grows directly with the number of path rules.

Final Time Complexity

Time Complexity: O(n)

This means the time to route a request grows linearly with the number of path rules in the Ingress.

Common Mistake

[X] Wrong: "Adding more path rules won't affect request routing time because it's just configuration."

[OK] Correct: Each incoming request is checked against all path rules until a match is found, so more rules mean more checks and longer routing time.

Interview Connect

Understanding how routing rules scale helps you design efficient Kubernetes Ingress setups and shows you can think about system performance beyond just writing configs.

Self-Check

"What if we changed the Ingress to use a single rule with a more complex path matching pattern? How would the time complexity change?"

Practice

1. What is the main purpose of using A/B testing with Kubernetes Ingress?

easy

A. To monitor CPU usage of pods

B. To increase the number of pods automatically

C. To split user traffic between different versions of an application

D. To backup Kubernetes cluster data

Solution

1. Step 1: Understand A/B testing concept in Kubernetes

   A/B testing with Ingress is used to route traffic between different app versions to test new features safely.

2. Step 2: Identify the purpose of Ingress in traffic management

   Ingress controls how external traffic reaches services, enabling traffic splitting for A/B testing.

3. Final Answer:

   To split user traffic between different versions of an application -> Option C

4. Quick Check:

   A/B testing = traffic split [OK]

Hint: A/B testing means splitting traffic between app versions [OK]

Common Mistakes:

• Confusing A/B testing with scaling pods
• Thinking Ingress is for monitoring only
• Assuming Ingress handles backups

2. Which annotation is commonly used in Kubernetes Ingress to split traffic by percentage for A/B testing?

easy

A. nginx.ingress.kubernetes.io/canary-weight

B. nginx.ingress.kubernetes.io/rewrite-target

C. nginx.ingress.kubernetes.io/ssl-redirect

D. nginx.ingress.kubernetes.io/proxy-body-size

Solution

1. Step 1: Identify annotations for traffic splitting

   The annotation 'nginx.ingress.kubernetes.io/canary-weight' is used to specify the percentage of traffic sent to the canary version.

2. Step 2: Differentiate from other annotations

   Other annotations like rewrite-target or ssl-redirect serve different purposes unrelated to traffic splitting.

3. Final Answer:

   nginx.ingress.kubernetes.io/canary-weight -> Option A

4. Quick Check:

   Canary weight = traffic percentage [OK]

Hint: Look for 'canary-weight' to split traffic by percent [OK]

Common Mistakes:

• Using rewrite-target for traffic splitting
• Confusing SSL redirect with traffic control
• Ignoring canary annotations

3. Given this Ingress snippet for A/B testing, what percentage of traffic goes to the canary service?

apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: example-ingress
  annotations:
    nginx.ingress.kubernetes.io/canary: "true"
    nginx.ingress.kubernetes.io/canary-weight: "30"
spec:
  rules:
  - host: example.com
    http:
      paths:
      - path: /
        pathType: Prefix
        backend:
          service:
            name: stable-service
            port:
              number: 80
      - path: /
        pathType: Prefix
        backend:
          service:
            name: canary-service
            port:
              number: 80

medium

A. 70%

B. 30%

C. 50%

D. 100%

Solution

1. Step 1: Read the canary-weight annotation

   The annotation 'nginx.ingress.kubernetes.io/canary-weight' is set to "30", meaning 30% of traffic goes to canary.

2. Step 2: Understand traffic split logic

   Traffic is split by percentage; 30% to canary-service, remaining 70% to stable-service.

3. Final Answer:

   30% -> Option B

4. Quick Check:

   Canary weight = 30% traffic [OK]

Hint: Check canary-weight value for traffic percent [OK]

Common Mistakes:

• Assuming canary gets 70% instead of 30%
• Confusing service names with traffic percentages
• Ignoring canary annotation

4. You configured this Ingress for A/B testing but all traffic goes to the stable service only. What is the likely error?

metadata:
  annotations:
    nginx.ingress.kubernetes.io/canary: "true"
    nginx.ingress.kubernetes.io/canary-weight: "50"
spec:
  rules:
  - host: test.com
    http:
      paths:
      - path: /
        pathType: Prefix
        backend:
          service:
            name: stable-service
            port:
              number: 80

medium

A. Missing canary backend path in spec

B. Incorrect canary-weight value format

C. Host name is invalid

D. Stable service port is wrong

Solution

1. Step 1: Check Ingress spec for canary backend

   The spec only has one backend for stable-service; no path defined for canary-service.

2. Step 2: Understand traffic routing requirements

   For canary traffic to work, a separate path with canary annotations and backend service must be defined.

3. Final Answer:

   Missing canary backend path in spec -> Option A

4. Quick Check:

   Canary needs separate backend path [OK]

Hint: Ensure canary backend path exists in Ingress spec [OK]

Common Mistakes:

• Assuming canary-weight alone routes traffic
• Ignoring missing canary backend path
• Blaming host or port without checking paths

5. You want to route 20% of users with header X-User-Type: beta to the canary service and the rest to stable. Which Ingress annotation setup achieves this?

hard

A. Use nginx.ingress.kubernetes.io/canary-by-cookie: "beta"

B. Use nginx.ingress.kubernetes.io/canary-weight: "20" only

C. Use nginx.ingress.kubernetes.io/canary: "false"

D. Use nginx.ingress.kubernetes.io/canary-by-header: "X-User-Type" and nginx.ingress.kubernetes.io/canary-by-header-value: "beta"

Solution

1. Step 1: Identify header-based routing annotations

   To route traffic based on header, use 'canary-by-header' and 'canary-by-header-value' annotations.

2. Step 2: Understand difference from weight-based routing

   Weight-based routing splits traffic by percentage regardless of headers; header-based routing targets specific users.

3. Final Answer:

   Use nginx.ingress.kubernetes.io/canary-by-header: "X-User-Type" and nginx.ingress.kubernetes.io/canary-by-header-value: "beta" -> Option D

4. Quick Check:

   Header-based routing uses canary-by-header annotations [OK]

Hint: Use canary-by-header and canary-by-header-value for header routing [OK]

Common Mistakes:

• Using only canary-weight for header routing
• Confusing cookie-based routing with header routing
• Setting canary to false disables routing