The stale bot is currently set up on the Amethyst repository. The original intention is probably to keep the issues “fresh”, in that they are still reproducible on the current code base.
Speaking of the stale bot will probably inevitably lead to the topic of our workflow or roadmap, but I consider that to be a separate thing and I’ll avoid discussing that as much as possible in this post.
In Amethyst I believe we have a separate roadmap (or something acting like that) from the issue tracker, sometimes in the form of RFC or working group goals. So IMO the issue tracker is largely for backlog, therefore we don’t need be too keen on keeping it short. I’m not saying that we can leave things intact indefinitely, since in some cases issues may actually go stale (or obsolete).
But the problem is that things are typically not stale. Given that we have limited (human) resources so development doesn’t happen so fast. The stale bot ends up just spamming notifications that don’t contribute to progress.
At last, an example of issue handling that I consider is working, is the Rust compiler repository. The workflow is pretty similar to us - issues sit just as they should do, while the majority work is led by working groups. The only thing being done to keep the issue tracker clean is triage - but it’s more like a bonus rather than mandatory process.