I recently right-sized the data and log files for an RDS database, leaving just a few GB of unallocated space. Shortly thereafter, something allocated, then deallocated, 1 GB of the margin of error. What might have caused this?

All of the databases - including tempdb - are either tiny (MB) or at exactly the size they've been set; they add up to 197 GB, leaving 3 GB free on a 200 GB instance. And yet:

Free space dipped from 3 to 1.5 GB, then returned

Clearly, if this happens again in a slightly greater magnitude, there will be a problem. For now, I can simply shrink one of the database files to give a larger margin, but I'd like to understand what actually happened.


AWS support came back with a comprehensive answer. In this case, it was transaction log backups, being written to the SQL disk before offloading to S3. There are a few more things which can consume space, mostly temporarily. As their tech outlined:

As a best Practice, I would recommend to keep at around 5% free space for similar processes (Log Backups, Error Log Files, Trace Files, etc), however this may vary based on high transactional volume or bulk data upload process , which may need high disk storage temporarily.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.