Six stages, from utilization baseline to ongoing review.
The same disciplined process whether it's twelve VMs or twelve hundred, and whether the goal is rightsizing an existing fleet or migrating a new one in.
Utilization Baseline
At least thirty days of CPU, memory, and disk telemetry collected per VM before any recommendation gets made — enough to capture weekly patterns and monthly peaks, not just a single snapshot.
Rightsizing Recommendations & Sign-off
Specific SKU-change recommendations per VM, with the reasoning behind each one, reviewed and approved by your team before anything is scheduled.
Scheduled Resize or Migration Execution
Resizing or migration happens during agreed maintenance windows, with a rollback plan defined before the change window opens, not improvised during it.
Commitment Strategy
Once the fleet is correctly sized, we layer in Reserved Instances, Azure Savings Plans, or Azure Hybrid Benefit against the stable, right-sized baseline — never against the inflated pre-rightsizing footprint.
Automation Setup
Auto-shutdown schedules for dev and test environments, automated patch windows through Azure Update Manager, and alerting for any VM that drifts back toward oversized.
Ongoing Monthly Review
A standing monthly review of utilization trends and spend, because a perfectly sized fleet today will not stay that way as workloads change — this is the stage that actually protects the savings.
Almost never the technical resize. Almost always the fear of the restart.
Resizing a VM in Azure typically takes minutes once a maintenance window is agreed. Getting stakeholders comfortable scheduling that window — especially for a VM nobody's touched in two years, where nobody's entirely sure what depends on it — is where real engagements take time. We treat that conversation as part of the actual work, not friction around it.