Examkin

Question

Difficulty: MediumCapacity Planning and Infrastructure Workload Scaling Optimization

A global gaming company is preparing for the launch of a new multiplayer game. The matchmaking service is deployed on Compute Engine Managed Instance Groups (MIGs) and communicates heavily over network sockets with minimal CPU overhead per connection. Historical load testing reveals that as concurrent player connections double, instance memory usage and active TCP socket connections saturate long before CPU utilization reaches 30%. What should you do to ensure the workload scales effectively during peak launch traffic while staying within Google Cloud resource limits?

  1. Configure the MIG autoscaling policy based on a custom Cloud Monitoring metric for active network connections, and verify/request regional CPU and instance quota increases in advance.Answer
  2. B
    Maintain the default MIG autoscaling policy based on target CPU utilization at 60%, and rely on dynamic auto-expansion of regional quota limits.
  3. C
    Migrate the matchmaking service to Google Kubernetes Engine (GKE) using Cloud Spanner for session state storage to automatically handle high-throughput network scaling.
  4. D
    Provision a Dedicated Interconnect connection to prevent compute instance capacity limits from blocking incoming player traffic.

Answer

Configure the MIG autoscaling policy based on a custom Cloud Monitoring metric for active network connections, and verify/request regional CPU and instance quota increases in advance.
For workloads where network sockets or memory exhaust before CPU load increases, autoscaling policies must target a custom Cloud Monitoring metric reflecting connection count. Additionally, effective capacity planning requires preemptive quota verification and increase requests before event launches to avoid resource allocation caps.

Step-by-Step Solution

1
Identify the primary workload scaling bottleneck.
The application saturates memory and TCP socket connections before CPU utilization rises significantly, making standard CPU-based autoscaling ineffective.
I/O-bound or connection-heavy workloads require custom metrics (such as active socket count or memory utilization) to trigger scaling events accurately.
2
Evaluate infrastructure scaling prerequisites and capacity limits.
Proactively reviewing regional GCP quotas for Compute Engine vCPUs and in-use IP addresses ensures autoscaling is not blocked during launch surges.
GCP quotas act as hard boundaries that require explicit request and approval in advance of planned capacity expansions.

Key Concept

Capacity Planning and Custom Metric Workload Scaling
Estimated Time:1m 30s
Rate this question

Topics

Capacity Planning and Infrastructure Workload Scaling Optimization