Dramatic cost increase for SQL Azure DB

Hello Robert, The Azure SQL Database cost has increased by approximately 60% in the most recent month, even though Database size was reduced from ~380 GB to ~204 GB and Usage remain limited/low.

Azure SQL Database pricing is not directly tied to actual data size or low usage. Instead, it primarily depends on Provisioned compute, Configured (max), Backup storage and retention and Service tier (General Purpose / Business Critical / Hyperscale).

Therefore, reducing data size alone does not guarantee cost reduction.

In Azure SQL:

• Compute is billed based on allocated capacity, not actual usage, Even with low activity, cost remains the same if the SKU/tier is unchanged. So, reduced usage does not reduce compute charges. vCore purchasing model - Azure SQL Database | Microsoft Learn
• Storage charges depend on provisioned or maximum configured size, not actual data stored. If max storage was not reduced, cost remains high.
• Backup storage is billed separately (especially in vCore model). Data deletions, index rebuilds, or high churn  in > crease backup size. This can increase cost even after shrinking data. Pricing - Azure SQL Database Single Database | Microsoft Azure.
• Costs may increase if any of the following were enabled or modified:
  • Geo-replication
  • Failover groups
  • Long-term retention backups
  • Advanced tiers (e.g., Hyperscale, In-Memory OLTP)

You can validate the cost by following these steps to identify the exact cause:

1. Check cost breakdown

• Go to Azure Portal > Cost Management > Cost Analysis
• Filter Service is Azure SQL Database
• Group by Meter / Resource.
• Identify whether spike is from Compute, Storage and Backup

1. Compare configuration (current vs previous month)

Validate:

• Service tier (GP / BC / Hyperscale)
• vCore or DTU level
• Pricing model (DTU vs vCore vs Serverless)

1. Validate storage settings: Check if Max Size (configured storage) changed and ensure it matches actual reduced requirement
2. If you are using a reserved instance, check if it is expired. Renew or switch to a savings plan.

Reference Documentation

• vCore vs DTU purchasing models
• Compute tiers (provisioned/serverless)
• Backup & long-term retention costs
• Cost analysis in Azure portal
• Azure Hybrid Benefit & reservations

Hope this helps! Let me know if you have any question on this, thanks.

Hello Robert, Just checking if the information shared was helpful. If so, please consider accepting the answer and giving it an upvote.

If you have any further questions or need additional clarification, feel free to reach out. Thanks!

Hello Robert, Just checking if the information shared was helpful. If so, please consider accepting the answer and giving it an upvote.

If you have any further questions or need additional clarification, feel free to reach out. Thanks!

Direct switching is indeed blocked unless you qualify for reverse migration.

• The database was migrated into Hyperscale (not created there).
• You are still within 45 days of that migration.

If either condition isn’t met, the only way out is a manual migration (bacpac export/import or other data movement tools). At ~204 GB, bacpac is technically supported but can be slow and error-prone. For larger DBs, failures are common.

Provisioned to Serverless (Hyperscale): Hyperscale does support serverless since GA in 2024. If the portal still shows “provisioned,” that’s usually a UI lag or metadata refresh issue, not a failed change. The definitive way to confirm is by querying sys.database_service_objectives:

SELECT name, service_objective
FROM sys.database_service_objectives
WHERE database_id = DB_ID('YourDatabaseName');

Interpretation:

• HS_S_Gen5_x > change succeeded, you’re on Hyperscale Serverless.
• HS_Gen5_x > change didn’t apply, resubmit the request.

How to Reverse Migrate a Database from Hyperscale - Azure SQL Database | Microsoft Learn

It did switch to 'serverless'. Will this reduce allocated storage?