Google Associate Cloud EngineerEnsuring successful operation of a cloud solutionHard
An e-commerce company is experiencing inconsistent application performance during peak sales events, leading to customer dissatisfaction. Their application runs on a Managed Instance Group (MIG) in Compute Engine. The team suspects that the instances are sometimes overwhelmed before new instances can be provisioned. They want to proactively scale up their application based on anticipated traffic patterns rather than reactively. Which scaling policy should they implement?
- AQueue-based autoscaling
- BLoad balancing utilization-based autoscaling
- CCPU utilization-based autoscaling
- DSchedule-based autoscaling
Show answer & explanationAnswer & explanation
Correct answer: D. Schedule-based autoscaling
Schedule-based autoscaling allows pre-configuring the MIG to scale up or down at specific times or intervals. This is ideal for anticipated traffic spikes (like peak sales events) as it ensures instances are ready proactively, preventing performance degradation and customer dissatisfaction.
Why the other options are wrong
- A. Queue-based autoscaling is typically used for asynchronous workloads (e.g., message queues) where processing backlog is the metric, not for synchronous web application performance during anticipated traffic spikes.
- B. Load balancing utilization-based autoscaling is also reactive, scaling based on the load balancer's capacity metrics, which again might not be proactive enough for anticipated events.
- C. CPU utilization-based autoscaling is reactive; it scales only after CPU usage is already high, which might be too late for sudden, anticipated spikes.
MIG Schedule-based Autoscaling
A feature of Managed Instance Groups (MIGs) that allows configuring scaling policies to increase or decrease the number of instances at predefined times or intervals.
- Enables proactive scaling for anticipated traffic patterns.
- Can be combined with other autoscaling signals (e.g., CPU) for dynamic adjustments.
- Helps ensure resources are ready before peak demand, improving application performance.
Memory trick: Schedule for Predictable Peaks.