Can you try redeploying your service?

Just done so, it seems to have dropped significantly, thanks. It looks like there's been some weirdness in the last few days, I don't know if they were related to your changes.

The volume dipped slightly just before it hit 800MB (pretty much right after I posted this). It dipped slightly more an hour ago but then breifly spiked to 2GB, with my deploy running out of memory at the same time despite my RAM usage remaining way below the cap (unless the email was wrong, but something caused a restart). The volume is at 173MB which I imagine is a normal amount of overhead?

We ran some cleanup commands on our end to reclaim the space. The root cause is file descriptors not being closed properly in your application. You'll need to fix this in your code - we won't be able to provide further assistance with this issue.

Interesting, I'll look into that, I wonder if my shutdown process isn't working properly or the SQLite driver is doing something weird with the connection pool. It's weird that redeploying normally doesn't fix it though. Thanks for your help