Hello fellow pants-wearers! I'm wondering where I can find documentation on

pantsd

, what it does and is responsible for. The only docs I can find are from pants v1.

Not sure about docs, but at a high level, its basically just a running process with an in-memory cache so you can get results even faster than the filesystem cache. You can turn it off with configs: https://www.pantsbuild.org/stable/reference/global-options#pantsd and it will impact performance, but not correctness.

Huh, I assumed we had explicit documentation about the daemon, but it looks like we don’t. Probably because `pantsd`*is* pants at this point.

FWIW, I'd like to see us get to the point where we're fast enough that

pantsd

is a "oh thats nice" rather than "goddamn I need this"

Thanks all 🙂 I'm working in a huge (275,000 files, 30-50 services) monorepo, and we're fighting resource constraints here and there. I find

pantsd

regularly eating some 4-6 gigs of RAM and

~/.cache/pants/

regularly gets bigger than 30-40GB. I'm mostly just curious to learn more about why.

You can get a slightly better idea of what’s going on with the stats subsystem: https://www.pantsbuild.org/stable/reference/subsystems/stats#memory_summary But yeah, usually what you’d find is cache on cache on cache - other than the downloaded and prepared libraries/dependencies/etc in your disk cache - and as you modify versions, that continually grows. I believe there is or will be some updates to pex that allows disk cleanup - but I haven’t looked into that deeply