Explanation
Why the waste happens and who it affects.
If an instance consistently consumes only a small portion of its allocated resources, it likely represents an opportunity to reduce costs through rightsizing. Without proactive reviews, these oversized instances can remain unnoticed and continue to incur unnecessary charges.
Billing model
The pricing dimensions that drive this cost.
Compute Engine VM instances are billed based on:
- Machine type and configuration - Cost scales with the number of vCPUs, memory, and any premium features (e.g., GPUs, local SSDs)
- Uptime - Charged per second, with a one-minute minimum
- Persistent disk and network usage
How to detect
5 checks to find it in your estate.
- Analyze average CPU and memory utilization of running Compute Engine instance
- Determine whether actual usage justifies the current machine type or custom configuration
- Review whether the workload could be met using a smaller predefined or custom machine type
- Check for constraints such as licensing, startup latency, or performance overhead that may require overprovisioning
- Validate with application or infrastructure teams whether the instance can be resized without impact
How to fix
How to remove the waste.
Resize the instance to a smaller predefined or custom machine type that better aligns with observed usage. For elastic workloads, consider deploying managed instance groups with autoscaling. Periodically review usage patterns to ensure continued alignment with resource needs.
Documentation
Vendor references for pricing and configuration.