RDS Blue/Green deployment does NOT allow shrinking

To successfully reduce storage using Blue/Green deployment, please follow these guidelines:

1. Storage Reduction Requirements:

• The target allocated storage must be at least 20% more than your current storage usage
• Calculate your required storage as follows:
  • Check your current storage usage via CloudWatch FreeStorageSpace metric
  • Add 20% to your current usage for the minimum target storage size
  • Example: If you use 7GB, minimum target size would be: 7GB + (7GB × 0.2) = 8.4GB

2. Important Considerations:

• Note that CloudWatch shows FreeStorageSpace in bytes while RDS Console shows storage in GiB
• The storage reduction only occurs if the above 20% requirement is met

3. Recommended Steps:

• Check your exact current storage usage using CloudWatch metrics
• Calculate the minimum target storage (current usage + 20%)
• Create the Blue/Green deployment with the calculated target storage
• Monitor the deployment process for any specific error messages

For additional assistance or if you continue to experience issues, please provide:

• The exact CloudWatch FreeStorageSpace metric value
• The target storage size you're attempting to set
• Any error messages received during the deployment

You are correct that the behavior you're experiencing with RDS Blue/Green deployment not allowing storage shrinking is unexpected and contrary to the intended functionality. This appears to be a limitation or potential bug in the RDS Blue/Green deployment process.

The Blue/Green deployment method is indeed one of the recommended approaches for reducing storage size in RDS instances, as it's supposed to allow you to create a new instance (Green) with a smaller storage size while keeping it in sync with your current instance (Blue). However, your experience suggests that this is not working as intended for your PostgreSQL 17.4 instance.

The fact that your new instance starts with 150 GB despite setting it to 40 GB during the Blue/Green deployment process is concerning. This behavior defeats one of the primary purposes of using Blue/Green deployments for storage optimization.

Given that your blue instance has never used more than 10 GB of data, and you're attempting to reduce it from 100 GB to 40 GB, this should theoretically be possible. The confirmation showing the new size as 40 GB, followed by the actual creation of a 150 GB instance, indicates a discrepancy in the deployment process.

This issue significantly impacts the ability to optimize costs and manage resources effectively, which is a core benefit of Blue/Green deployments for RDS.

Since this appears to be a potential bug or limitation in the RDS Blue/Green deployment process, it would be advisable to:

1. Document your steps and observations in detail.
2. Open a support case with AWS to report this issue. Provide them with the specifics of your deployment attempts, including screenshots and any relevant logs.
3. Request clarification on whether this is a known limitation or a bug that needs to be addressed.

In the meantime, you might need to consider alternative approaches to manage your storage, such as:

1. Manually creating a new RDS instance with the desired lower storage size and migrating your data.
2. Implementing more aggressive data archiving strategies to minimize storage usage.
3. Exploring options for optimizing your current database to reduce storage consumption, even if you can't reduce the allocated size.

It's important to bring this to AWS's attention as it could affect other users and contradicts the expected behavior of Blue/Green deployments for RDS storage management.
Sources
Reduce the RDS MS SQL Storage post data archival | AWS re:Post
Reduce Storage Size - Oracle RDS Database | AWS re:Post