I have a service with a 5gb volume mounted. According to the observability metrics, the amount of data on the volume fairly recently jumped from just-under 200mb to almost 500mb. There was some activity on the app during that window, but quite minimal, and definitely didn't involve 200+ mb of data being written to the volume.

Indeed, when I use railway ssh to log into the box and look at the data on there, it's a much smaller amount (131mb). I know there's some overhead associated with the volume, and I assume that's the difference between 131mb of visible data and the original ~190mb that i saw in metrics, but I'm concerned that it seems like almost 300mb of additional overhead was added, and that seems like a lot of overhead.

I'm concerned about my ability to accurately monitor and predict volume usage if the amount of overhead is so large and variable.

The jump in usage may have correlated with when I increased the volume size, but I'm not positive about that.