Include downgrade timing and tooling in a Premium SSD tier-change plan

Source facts

Azure Premium SSD managed disks can change performance tier without resizing the disk. A tier cannot fall below the disk’s baseline, and downgrades are limited to once per twelve hours. Shared-disk changes require stopping all attached VMs. Microsoft additionally warns that Terraform shuts down a running VM when used for this change, unlike the other documented methods. Microsoft Learn.

Applicability

Review a Premium SSD managed disk, not Ultra Disk or Premium SSD v2 performance settings. Identify its baseline and current tier, the supported target, and every attached VM. Ask the workload owner to state how long the higher performance is required before deciding when to return.

DSE recommendation

Record the baseline tier, permitted return time, attachment scope, and execution tool before approving a temporary tier increase. Make the change plan explicit about any required shutdown rather than treating the feature’s general online capability as a property of every tool or attachment arrangement. Assign an owner to revisit the temporary increase after the workload event ends.

Verification

Inspect the resulting tier and unchanged capacity after the approved operation. Run the agreed workload measurement and compare it with the purpose of the increase. Before returning to baseline, check the downgrade interval and the same attachment/tooling conditions again. Preserve the original tier, actual change times, and final result so a failed or delayed return is visible instead of becoming an undocumented permanent setting.

Official references

Microsoft Learn: Performance tiers for Azure managed disks. Source reviewed September 9, 2026.