Skip to content
Costframe

Turn a cost signal into an engineering review.

Find the resource behind a charge, inspect its utilization and prepare the owner's next step. Review dependencies and workload needs before testing a change.

Explore a sample review
An engineering notebook beside a layered workload model, ready for inspection.

Prepare a change your team can test

A low-utilization finding starts an investigation, not a resize. Agree the workload's peak needs and service targets, then prepare a bounded test with a rollback path.

  • Check current configuration and collected evidence.
  • Confirm missing signals and business requirements.
  • Record dependencies, safe testing and rollback.
Implementation review guidance
Azure resource example · sample dataA CPU signal starts the review.
Sandbox / rg-developmentvm-build-07

Virtual machine · West Europe

22%Peak in the public CPU sample

14 daily readings · not a capacity guarantee

Read all 14 CPU samples
Illustrative daily CPU readings, percent
DayCPU
16%
211%
38%
47%
516%
610%
78%
85%
97%
1022%
1112%
128%
136%
149%
Before a sizing decision
Available in this sample
  • Resource and billing context
  • 14-day CPU sample
  • Configured team: Developer experience
Not collected
  • Memory utilization
  • Disk IOPS and performance
Owner review still needed

Confirm peak build needs, dependencies, a safe test and rollback before changing capacity.

Existing public sample, not customer evidence. Low CPU alone does not prove spare capacity. Your team controls any external change.

After the test, check service performance first, then matching-period usage and billed cost.

Work through a VM sizing decision

Your starting point

A cost increase without a clear resource, owner or safe implementation decision.

What Costframe adds

Resource-level charges, available configuration, retained evidence and reviewable findings.

The next team discussion

The evidence window, missing telemetry, owner requirements and a proposed test/rollback plan.

Check workload constraints before choosing capacity.

CPU alone is not proof of spare capacity. Memory, IOPS, dependencies and peak requirements need review. Supported guidance is non-executing; it does not create Terraform/CLI commands or modify customer resources.

Azure is generally available. AWS is a private pilot and Google Cloud a private preview. Evidence coverage, native currency and permissions remain explicit; your team implements changes outside Costframe.

Provider scope and setup

Review one workload before planning a change.

Bring its evidence window, dependencies and rollback criteria to the owning team.

Start with Costframe