An Azure backup service that provides built-in management at scale.
Short answer: I couldn't find any documented self-service or request-based way to purge a soft-deleted backup instance early, and the docs say soft delete can't be disabled for Backup vaults in Australia East (secure by default is GA there). So the practical route is to make the IaC work with the instance you already have, or wait out the retention period.
What the docs say
- A deleted backup instance goes into a soft-deleted state and is kept for the soft-delete retention period (14 days by default, configurable from 14 to 180 days). After that it is deleted permanently.
- You can't disable soft delete where secure by default is in preview or GA, and Australia East is listed as GA for Backup vaults.
- A soft-deleted instance can be recovered ("Undelete"). It comes back as Stop protection with retain data, and backups don't resume until you select Resume backup.
I didn't find a purge/skip-retention operation in that documentation, so I wouldn't count on one. If you need an exception, the only route I'm aware of is a support request, and I can't say whether it would be granted.
Options
- Align the IaC with the existing instance. If the name is what clashes, check what the existing instance is called and point the template at that name instead of generating a new one. You can list it with:
az dataprotection backup-instance list -g <rg> --vault-name <vault> \
--query "[].{name:name, state:properties.currentProtectionState}" -o table
- If the instance is soft-deleted: undelete it (Backup center / Resiliency > Backup instances, filter Protection status == Soft deleted > Undelete), then resume backup or stop protection as needed, and have the pipeline target that instance rather than creating a new one.
- Wait for retention to expire, then deploy. Note the retention that was active at the time of deletion governs it, so lowering it now won't shorten an already-deleted item.
- Deploy to a different vault if you need the pipeline unblocked today.
Please double-check which of these matches your actual error (a name conflict vs. an item still in soft-deleted state), because the fix is different. If you can paste the deployment error, I can narrow it down.
Docs: