Google Associate Cloud EngineerEnsuring successful operation of a cloud solutionHard
A company has a critical application running on Compute Engine instances that requires strict uptime. They need a mechanism to automatically restart instances that become unhealthy due to application-level issues, even if the underlying VM is still running. Which feature should be configured to address this requirement?
- AAutomatic restarts on the instance configuration
- BCloud Monitoring Alert Policy with Autohealing
- CInstance Scheduler
- DManaged Instance Group (MIG) with an HTTP Health Check
Show answer & explanationAnswer & explanation
Correct answer: D. Managed Instance Group (MIG) with an HTTP Health Check
A Managed Instance Group (MIG) can be configured with an HTTP Health Check that monitors the application's responsiveness. If an instance fails the health check, the MIG will automatically replace or restart the unhealthy instance, ensuring application availability.
Why the other options are wrong
- A. Automatic restarts configured directly on a standalone Compute Engine instance only address VM failures (e.g., host maintenance), not application-level unhealthiness where the OS is still running.
- B. Cloud Monitoring Alert Policy can notify about issues, but it doesn't inherently restart instances for application-level unhealthiness; it would typically trigger a notification, and an external mechanism or a MIG's autohealing would perform the action.
- C. Instance Scheduler is used to start and stop instances on a predefined schedule, not to react to instance unhealthiness or application failures.
MIG Autohealing
A feature of Managed Instance Groups (MIGs) that automatically replaces or restarts unhealthy instances based on configured health checks, ensuring application availability.
- Requires a Managed Instance Group
- Uses health checks (HTTP, TCP, SSL) to determine instance health
- Automatically replaces or restarts unhealthy instances
- Ensures application-level availability, not just VM uptime
Memory trick: MIGs are like doctors for your VMs, always checking heartbeats.