Daniel Swarbrick

@SuperQ I forked prometheus/procfs recently, and had an initial poke around. My WIP is in https://github.com/dswarbrick/procfs/blob/mdraid-sysfs/sysfs/mdraid.go. The code is in its early stages - I intend to split up the...

@frittentheke Sadly I let this slip off my radar for a long time. There was also some discussion over the "right" way to implement reading of a large number of...

@jooola I don't have the exact link off the top of my head, but it mostly revolved around a couple of fragile approaches. One involved basically reading every file in...

@jooola The kernel md admin guide describes how many of the md-related sysfs (and `/proc/mdstat`, IIRC) support select/poll, to efficiently watch for changes. I don't know how useful this is...

Linking to PR #404 which introduced this regression. The crux of the issue is that the shipped testdata fixtures will overflow ints in the ProcStat struct on 32-bit archs.

In terms of fixing this, we ran into a pretty similar situation before, which I fixed in #314, i.e. split the tests into 32-bit and 64-bit variants, using Go build...

I strongly suggest that future enhancements to mdraid stats use the sysfs entries, rather than add any more parsing of /proc/mdstat. The sysfs interface is designed to be machine-readable, whereas...

There seems to be some confusion about how to expose a "good" Prometheus metric in this library. Take this example: ``` # HELP nodejs_active_handles Number of active libuv handles grouped...

The certificates were regenerated in [v0.36.0](https://github.com/prometheus/common/releases/tag/v0.36.0) so this should no longer be an issue.

Wouldn't it be a lot simpler to just fix the PATH env var in the cronjob?