Storage on Azure Database for PostgreSQL flexible server can only be increased, never decreased. Both manual and automatic approaches are supported, but all rely on the platform allowing scaling operations in the selected region and SKU.
From the documented behavior:
- Storage can be scaled up for a flexible server using the Azure portal under Compute + storage; for Premium SSD, the storage size is chosen from a predefined list of sizes (each value double the previous, minimum 32 GiB). For Premium SSD v2, any size in GiB can be entered directly, minimum 32 GiB.
- Decreasing storage is not supported in any case.
- For Premium SSD:
- Increasing from 32 GiB–4 TiB to another size in that range, or from 8 TiB–32 TiB within that range, is online and does not require downtime.
- Increasing from ≤ 4,096 GiB to > 4,096 GiB requires a server restart and disables disk caching.
- For Premium SSD v2:
- Any increase in disk size always requires a server restart.
- Host caching is not supported.
The errors described (InternalServerError and “location is restricted for provisioning”) are not covered in the scaling documentation and indicate a platform or regional restriction rather than a configuration issue on the server itself. The documentation explicitly notes that Premium SSD v2 is in preview and that regional and quota limits can apply; it also states that Premium SSD v2 offers up to 32 TiB per region per subscription by default, with higher capacity available only by quota increase request.
Given that:
- The server is nearly out of space and storage must be increased (storage can only grow, not shrink).
- The portal operation to increase storage is failing with platform-level errors related to provisioning in the region.
- The issue involves regional restrictions and internal errors that are not resolvable by configuration changes described in the public scaling guidance.
The only viable path is to engage Azure Support with the provided tracking IDs so the platform team can:
- Investigate the InternalServerError and “location is restricted for provisioning” conditions for the specific region and server.
- Confirm whether there is a regional restriction, capacity issue, or preview/feature limitation blocking scaling in East US for the current SKU.
- Either unblock storage scaling on the existing server or advise and assist with a supported migration path (for example, to another region or to a different storage type/tier that supports scaling in that region).
While waiting for support, monitor storage usage closely using the storage_percentage or storage_used metrics and keep emergency cleanup measures in place. Consider enabling storage autogrow if available for the current storage type, so the platform can automatically increase storage before the disk becomes full.
References: