Examkin

Question

Difficulty: HardManaging Compute Engine Resources

An organization runs a high-throughput backend service on a Compute Engine Managed Instance Group (MIG) configured with CPU utilization autoscaling. During peak traffic events, newly provisioned VM instances require approximately 5 minutes to start application dependencies and warm up local data caches. During this warm-up phase, CPU usage spikes temporarily, causing the autoscaler to continuously launch redundant VM instances before the existing new instances can handle traffic. Which action should the Cloud Engineer take to prevent premature and excessive scaling during instance startup?

  1. Configure an appropriate cool-down period in the MIG autoscaling policy to allow instances to fully initialize before their metrics are evaluated.Answer
  2. B
    Increase the target CPU utilization threshold in the autoscaling policy to 100% so initialization spikes are ignored.
  3. C
    Reconfigure the instance template to use Spot VMs, which automatically pause metric collection during instance boot.
  4. D
    Deploy additional MIGs in multiple secondary regions to bypass regional compute instance quotas during traffic bursts.

Answer

Configure an appropriate cool-down period in the MIG autoscaling policy to allow newly created instances to complete their application startup before their CPU metrics are included in scaling decisions.
Configuring a cool-down period (also known as the initialization period) in the Managed Instance Group autoscaling policy specifies how long the autoscaler must wait after an instance boots before incorporating its usage metrics into autoscaling decisions. By setting this duration to cover the 5-minute startup and cache-warming window, metrics generated during initialization are ignored, preventing continuous, redundant instance provisioning.

Step-by-Step Solution

1
Analyze the cause of premature autoscaling during instance startup.
Identified that high CPU usage during application boot and cache warming causes the autoscaler to misinterpret initialization load as sustained demand.
Without a delay, metrics collected during startup trigger immediate, unnecessary scale-out events.
2
Identify the Google Cloud recommended mechanism for handling VM initialization timing in MIGs.
The cool-down period (initialization period) parameter defines how long the autoscaler waits before assessing telemetry from a new instance.
Setting the cool-down period to match or exceed the 5-minute boot time ensures new instances reach steady-state operation before contributing to autoscaling calculations.
3
Evaluate the autoscaler configuration change.
Configuring the cool-down period stabilizes the instance count during scale-out events without distorting baseline target utilization thresholds.
This maintains appropriate autoscaling responsiveness while preventing instance thrashing.

Key Concept

Compute Engine MIG Autoscaler Cool-Down Period Configuration
Estimated Time:2m 0s
Rate this question

Topics

Managing Compute Engine Resources

More questions from this topic