It's been a few months since I last asked (somwhere, can't find it) - how healthy does our repo look to you?
I think it looks well-organized, active, and friendly :)
The only thing that stands out to me as a yellow flag is our open issue count.
Issues: 664 open / 781 closed :thinking:
PRs: 43 open / 2,773 closed :smiley:
Time-wise, here's when those open issues are from:
0 open issues from 2019 or earlier
65 open issues from 2020
239 open issues from 2021
371 open issues from 2022
What do you think about these metrics - is that too many open issues? Is it intimidating to newcomers, or looking unmaintained? We definitely keep old good issues around for project management purposes, which is valid.
The "appearance of maintenance" aspect is more important in the future when we come out of "work in progress" phase into "v0.1" phase, but I'm just curious to hear opinions on its current state.
(also, completely coincidentally, our open issue count is equal to the issue # for one of our biggest open issues, and that makes me very happy haha)
Screenshot-from-2022-10-08-19-41-38.png
Lots of popular open source projects have a massive amount of open issues so that's not something I'd be worried about. I often don't create an issue when I implement or fix things, we could try to do that more to improve the open/closed ratio.
Whenever I’m trying to determine liveness of a project I look at the time of the last couple commits. Issues don’t bother me at all unless it’s obvious that things are never ever being closed out (e.g. Quill)
it seems like we have a growing community and a corresponding growing issue count. Fine with me!
Great perspectives, sounds like our issue count/ratio is fine!
For PRs & branches I've been tidying/reviving when they get older than 6-9months
Do we want a soft limit on issue age? Maybe 2 years?
what problem would that solve?
if we didn’t address an issue in two years we’d just get people saying “hey, what was the solution here?”
if the answer is “we got tired of it being open”, I think that’d be a worse signal than a big open count
if stuff is truly stale we can have some bot take care of pings and closing. There are a few to choose from, IIRC
but I kinda think it’d be putting the cat in front of the horse at this stage
(haaaaaah what a good typo. I love it and I’m leaving it. On mobile this morning.)
There's already a workflow for stale PRs, so one for stale issues wouldn't be a stretch.
I think it is good to ping old issues at some sort of cadence. Generally there are three reasons an issue is old:
1) something that is future planned work but with a long time horizon
2) something obsolete that never got closed
3) an issue were we are waiting on a reply from someone and can't continue work without it.
For 2 and 3, it would be good to ping stale issue and eventually auto close them.
For 1, it is fine if the issue gets pinged (though a tad annoying if it bloats the thread), but we definitely don't want them to close.
We can have a bot ignore issues with certain characteristics like label. That way we can differentiate 1) from 2) and 3).
Last updated: Jun 16 2026 at 16:19 UTC