Google Professional Cloud Developer Question 296
Select 3Google Cloud PlatformYou are developing a distributed application running on Google Kubernetes Engine (GKE) that uses multiple microservices. To ensure the application's resilience, you decide to implement chaos engineering principles by simulating failure scenarios. Which of the following actions are valid approaches to test the system's fault tolerance using chaos engineering?
- A
Introduce random delays in API responses between microservices to test timeout handling.
- B
Terminate a random pod in the cluster to observe how the application recovers.
- C
Manually scale up all pods to their maximum limit to test scalability under high load.
- D
Introduce network partitioning to simulate communication failures between microservices.
- E
Write custom scripts to restart the Kubernetes cluster entirely to simulate complete system failure.
Show answer and explanation
Correct answers: A, B, D
Explanation
Chaos engineering involves introducing controlled failures into a system to test its resilience and fault tolerance. Valid practices include simulating latency, killing pods, and introducing network issues. These practices help uncover weaknesses and improve the overall stability of the system. Extreme or unrelated actions, such as full cluster restarts or performance scalability tests, are not part of chaos engineering.
- A. Correct.
Introducing random delays simulates latency issues, which helps test the application's ability to handle timeouts and degrade gracefully. This aligns with chaos engineering principles.
- B. Correct.
Terminating a random pod simulates unexpected failures, allowing you to observe how the application responds and maintains availability. This is a core practice in chaos engineering.
- C. Incorrect.
Scaling up pods to their maximum limit is a scalability test, not a chaos engineering practice, as it focuses on performance under load rather than fault tolerance.
- D. Correct.
Introducing network partitioning simulates communication failures, helping you evaluate the system's resilience to such scenarios. This is a valid chaos engineering approach.
- E. Incorrect.
Restarting the entire Kubernetes cluster is an extreme scenario and not aligned with typical chaos engineering practices, which focus on smaller, controlled experiments to test resilience.