3.3 Mitigating Incident Impact on Users PROFESSIONAL-CLOUD-DEVOPS-ENGINEER Practice Questions (Page 3)

Questions 11–15

1. 11foundation · easy

   A regional outage is affecting a service in us-central1. The service is also deployed in us-east1. Which Google Cloud load balancing approach can automatically redirect user traffic to the healthy region?

   • AUpdate the DNS A record to point to the us-east1 instance's IP address.
   • BConfigure a global external HTTP(S) load balancer with a backend service in each region and use health checks to fail over.
   • CCreate a new firewall rule to allow traffic only to us-east1.
   • DUse Cloud CDN to cache all content and serve from edge locations.
   Reveal Solution Discussion Select an answer first
2. 12application · medium

   A video streaming platform is performing maintenance on a specific backend instance group. They need to take these instances out of service for a hardware refresh without dropping any active user video streams. Users are connected via long-lived WebSocket connections. What is the correct approach to achieve this?

   • AEnable connection draining on the instance group, wait for active connections to close, then terminate the instances.
   • BImmediately delete the instances from the instance group.
   • CReduce the max instances to zero in the autoscaler.
   • DStop the instances without draining connections.
   Reveal Solution Discussion Select an answer first
4. 13expert · medium

   A team is deploying a new version of their application. They have a feature flag that controls the new feature. They want to use a canary release to test the new feature with a small percentage of users. They have set up the feature flag to be on for 5% of users. After monitoring for 30 minutes, they see no issues. What is the best next step?

   • ARoll back the feature flag to off for all users.
   • BRemove the feature flag from the codebase.
   • CDeploy the feature to all users immediately.
   • DIncrease the feature flag percentage to 25%.
   Reveal Solution Discussion Select an answer first
5. 14expert · hard

   A company runs a stateful application on a managed instance group. They need to perform a rolling update of the application. The application requires that connections are drained gracefully before an instance is terminated. The team has configured a health check and set the 'draining timeout' to 60 seconds. However, during the update, some users are experiencing connection resets. What is the most likely cause?

   • AThe draining timeout is too short for the application to complete in-flight requests.
   • BThe health check is failing on the instances being updated.
   • CThe instance group is not configured for connection draining.
   • DThe autoscaler is scaling down instances too aggressively.
   Reveal Solution Discussion Select an answer first
6. 15application · medium

   A team has deployed a new version of their application to a small subset of users using a canary release. The canary is showing a slight increase in error rates compared to the stable version, but it is not critical. The team wants to continue monitoring without affecting the majority of users. What should they do?

   • APromote the canary to full production.
   • BRoll back the canary immediately.
   • CIncrease the canary percentage to get more data.
   • DKeep the canary at the current percentage and monitor for a longer period.
   Reveal Solution Discussion Select an answer first