Prevent unintended default-profile gaps between autoscale schedules

Source facts

Autoscale checks fixed-date profiles before recurring profiles, then the default; only the first applicable profile supplies rules and capacity limits. In the portal, explicitly adjoining end and start times can introduce a one-minute default-profile interval. Microsoft advises leaving the end time blank for contiguous profiles. Microsoft Learn.

A recurring profile without an end time is supported through the portal and ARM templates. Omitting the CLI end time instead defaults it to 23:59. PowerShell requires a separately scheduled copy of the default profile to end a recurrence. Microsoft Learn.

Applicability

Use this review for Azure autoscale settings with multiple scheduled profiles. Distinguish a deliberately scheduled return to default from an intended direct transition between recurring profiles. Keep the authoring tool in scope because the documented behaviors differ.

DSE recommendation

DSE recommends a boundary table listing the intended active profile immediately before and after each transition. Include its capacity limits and any fixed-date override. Review the deployed profile definitions, not only the schedule as described in a change ticket. Ask the workload owner to approve any interval governed by the default profile, even if brief.

Verification

Rehearse an accelerated schedule on a disposable target. Observe each transition and compare it with the boundary table, especially the minute between adjacent profiles. Inspect the exported settings for unexpected default copies or end times. Retain the observed active profile and instance count without treating one rehearsal as proof of future capacity sufficiency.

Official references

Microsoft Learn: Autoscale with multiple profiles.