GCP Associate Cloud Engineer Practice Question 36: MIG Autohealing Health Check
Google Cloud Associate Cloud Engineer · Test 4 · question 6 of 10
In a managed instance group, a VM sometimes keeps running while the web server process on it hangs and stops answering requests. The load balancer already stops sending traffic to the VM, but the VM stays in the group indefinitely and capacity is lost. You want unhealthy VMs to be replaced automatically.
What should you do?
Show answer and explanation
Answer:
Autohealing uses an application level health check (for example, HTTP on the serving port) and recreates any instance that fails it. It is configured on the managed instance group itself, so the VM is replaced automatically.
Option A only affects how quickly the load balancer detects a problem; load balancer health checks control traffic routing and never recreate VMs. Option C keeps extra capacity but leaves the broken VM in place. Option D depends on a person, which is the manual step you want to remove.
Reference: About repairing VMs for high availability (autohealing)