How to
How to configure customer-controlled VM scheduling
Define exact VM scope and schedule windows controlled by your team, not by automatic environment assumptions.
Back to How To GuidesPolicy Model
- Schedule policy stores timezone, start/stop window, weekend behavior, and `selectedResourceIds`.
- Only selected VMs are in scope for scheduling impact analysis.
- This model is tenant-specific and fully auditable.
- Policy changes should be reviewed like any other cost automation because they can affect availability.
Prerequisites
Fresh inventory
Run a scan before creating the policy so the VM list reflects current resources.
Clear ownership
Confirm who owns each selected VM and who can approve schedule changes.
Write permissions
Read-only cloud credentials can estimate impact, but cannot execute start/stop actions.
Business window
Document timezone, working hours, weekend behavior and holiday expectations.
Cloud Permissions
Basic VM scheduling uses provider control-plane APIs. TurboFinOps does not install an agent on the VM for start/stop execution. Guest agents may still matter for provider-specific OS features, scripts or graceful in-guest operations, but the scheduling action itself depends on cloud IAM/RBAC.
AWS EC2
Use describe permissions for discovery and start/stop permissions for execution.
ec2:StartInstancesec2:StopInstancesec2:DescribeInstancesAzure VM
TurboFinOps deallocates VMs for real compute savings, so deallocate permission is required.
Microsoft.Compute/virtualMachines/start/actionMicrosoft.Compute/virtualMachines/deallocate/actionMicrosoft.Compute/virtualMachines/readGCP Compute
Grant these on the project or a narrower custom role used by the connected service account.
compute.instances.startcompute.instances.stopcompute.instances.getcompute.instances.listSetup Steps
- 1. Open `Cost Savings > Scheduling`.
- 2. Set policy name, timezone, start time, stop time, and weekend switch.
- 3. Search and select exact VMs in the scope table.
- 4. Save policy and refresh view.
- 5. Review policy history after save and confirm the intended resources remain selected.
Impact Review
- Use `Selected VM Impact` table to inspect estimated savings per VM.
- Validate business-critical workloads are excluded from selection.
- Re-check total projected monthly savings before execution rollout.
- Compare projected savings against business risk; a small saving is rarely worth production interruption.
Exceptions
Exclude production databases, always-on bastion hosts, shared build workers and any VM with unclear ownership until the owner signs off. If a VM needs temporary exclusion, record the reason in the action or policy notes so Finance and Operations understand why projected savings changed.
Operational Notes
Start/stop automation pipeline can be enabled progressively after policy review. Keep a rollback playbook for workloads with strict availability requirements. Start with non-production scopes, watch the first week of execution history, then expand only after owners confirm there were no missed business-hour starts.
TurboFinOps
Start with one cloud scope. Prove savings fast.
Connect AWS, Azure, or GCP and get actionable findings, score trends, and auditable remediation paths in minutes.