How to successfully reduce RDS storage size using Blue/Green deployment?

Question

I’m trying to reduce allocated storage for an RDS MySQL instance using the Blue/Green deployment approach, following this AWS blog:

[Shrink storage volumes for your RDS databases and optimize your infrastructure costs](https://aws.amazon.com/blogs/database/shrink-storage-volumes-for-your-rds-databases-and-optimize-your-infrastructure-costs/)

Source DB details
- Engine: MySQL 8.0.42
- Current allocated storage: 330 GB
- FreeStorageSpace from CloudWatch: ~270 GB
- Storage autoscaling: Disabled
- File system configuration: Upgrade available

According to the blog, the new allocated size must be at least 20% greater than FreeStorageSpace.
So in my case:
```
330 GB (allocated) - 270 GB (free) = 60 GB used
60 GB * 1.2 = 72 GB minimum
```
I set the Blue/Green target storage to 80 GB, which should be above the required threshold.

However, the deployment fails with the following logs:
```
September 04, 2025, 10:24 (UTC+02:00) 
Your blue/green deployment bgd-hqavqgeih97lny10 will create a read replica of dltdb with storage type gp3, and allocated storage 80.

September 04, 2025, 11:43 (UTC+02:00) 
Failed to reduce storage for blue/green deployment bgd-hqavqgeih97lny10. 
See the documentation for valid target-allocated-storage values.
```

My questions:
- Why would 80 GB not be considered a “valid target-allocated-storage value” when it satisfies the +20% storage allocation rule from the blog?
- Are there additional hidden constraints (minimum absolute size, step granularity, engine-specific requirements, etc.) that I might be missing?
- Has anyone successfully reduced RDS storage with this method for MySQL, and what target values worked for you?
- Are there any logs I can check that would provide more descriptive details about why the storage reduction failed?

Any insights or examples would be greatly appreciated.

Jeremy · Accepted Answer

I managed to solve this.

I was able to reduce the size of the storage by allocating **100 GB**, which is about **+65% of the used space** according to CloudWatch. That deployment completed successfully.

I’m still not sure if it would have worked with a value somewhere between the previously tested **80 GB** (already +35% of used space) and the working **100 GB**.

At one point I suspected the extra space requirement might be related to the **file system configuration upgrade** step mentioned in the docs/blog. To test this, I tried creating a **read replica with the current filesystem**, promoted it, and then attempted to reduce the storage to **80 GB**. That failed with the same *“valid target-allocated-storage values”* error.

So in practice, the lower bound seems to be **stricter than the blog’s +20% rule**, at least for my instance.

**Takeaway:** If you run into this issue, don’t just aim for the minimum +20%. In my case, I needed closer to **+65% of used space** before the Blue/Green deployment succeeded. It may be worth experimenting with higher values if the lower bound fails.

re:Post Agent · Answer

To successfully reduce RDS storage size using Blue/Green deployment, you need to follow specific guidelines and understand certain limitations:

## Steps for RDS Storage Reduction via Blue/Green Deployment

1. **Create a Blue/Green Deployment** through the Amazon RDS console:
- Navigate to the storage section
- Set the allocated storage for the green database to your desired smaller size
- For example, if your blue database is 400GB, you could create a green database with 200GB

2. **Wait for Deployment Completion**:
- The creation process may take hours or days depending on database size, storage type, and IOPS
- Monitor the progress until the green environment shows as available

3. **Perform the Switchover**:
- Once the green environment is fully deployed and available, orchestrate a managed switchover
- The promotion process is quick and can be completed in as little as a minute with no data loss

4. **Clean Up**:
- After successful switchover, delete the blue/green deployment
- Delete the old blue instance to save on costs

## Regarding Your Specific Issue

Your calculation approach is correct in principle, but there might be several reasons for the failure:

1. **Engine Support**: Ensure your MySQL version is fully supported. Blue/Green Deployments with storage shrinking is available for MySQL major versions 5.7 and higher.

2. **Hidden Constraints**: There might be additional factors beyond the 20% rule:
- The system might be calculating used space differently than your CloudWatch metrics suggest
- There could be temporary storage usage not reflected in the FreeStorageSpace metric
- MySQL might have specific storage requirements based on your database configuration

3. **Storage Optimization**: Before attempting the Blue/Green deployment:
- Consider running OPTIMIZE TABLE operations on large tables
- This can help reduce fragmentation and ensure more accurate space calculations

4. **Alternative Approach**: If Blue/Green deployment continues to fail, you could consider:
- Creating a new RDS instance with lower allocated storage
- Using AWS DMS Serverless to migrate data from your current database
- This approach has been successfully used for significant storage reductions

Remember that the Blue/Green deployment method for storage reduction is relatively new (announced in November 2024), so you might be encountering edge cases in the implementation.  
**Sources**  
[Amazon RDS Blue/Green Deployments support storage volume shrink - AWS](https://aws.amazon.com/about-aws/whats-new/2024/11/amazon-rds-blue-green-deployments-storage-volume-shrink/)  
[Shrink storage volumes for your RDS databases and optimize your infrastructure costs | AWS Database Blog](https://aws.amazon.com/blogs/database/shrink-storage-volumes-for-your-rds-databases-and-optimize-your-infrastructure-costs/)  
[Scaling Amazon RDS for MySQL performance for Careem’s digital platform on AWS | AWS Database Blog](https://aws.amazon.com/blogs/database/scaling-amazon-rds-for-mysql-performance-for-careems-digital-platform-on-aws/)  
[Managing capacity automatically with Amazon RDS storage autoscaling - Amazon Relational Database Service](https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/USER_PIOPS.Autoscaling.html)

How to successfully reduce RDS storage size using Blue/Green deployment?

Steps for RDS Storage Reduction via Blue/Green Deployment

Regarding Your Specific Issue

Add your answer

Relevant content