How do you make sure your staging server doesn't get choked by invisible cache files?

There is nothing quite like getting an alert at midnight that your server disk hit 100% capacity because an image optimization script was silently hoarding cache files in the background.

We missed deleting the working directory after processing user uploads and it completely paralyzed our web traffic until we manually cleared the drive.

For teams building microservices that handle heavy file processing what automated safeguards do you use to catch runaway storage usage before it crashes everything?