-
-
Notifications
You must be signed in to change notification settings - Fork 2.1k
support prometheus_client >= 0.4 #4221
Comments
I see NixOS having issues with this, Gentoo also dropped 0.3.1 from the repo about a week ago. |
If we start fresh, is synapse compatible with >0.4.0 ? |
Not entirely. The tests fail: #4167 |
Sad, I think this is my last hardblocker for installing it on my system (gentoo). I don’t really intend on porting <prometheus-client-0.4.0 :/ |
@nado , I have been running with the following as
which seems to work, but then again it shouldn't really matter if you don't use the metrics functionality. |
Yeah I could, but since I intended to make a PR against gentoo tree as a long goal, it seems weird to add back a removed ebuild. It might be okay. But there’s one thing I don't understand, is it a required or an optional dependency? |
It is required as specified by the build system. I just mean that if you don't use those features, it will never be used at runtime. IMO an ebuild should reflect what upstream says, and only make functionality optional at such time upstream is willing to guarantee that it is written sufficiently modularly. I have also been running synapse quite well on both x86_64 and armv7a with this ebuild coupled with the gerislay overlay. This discussion is probably getting a bit off-topic now so if there is another more Gentoo-focused location (issue on an overlay repo?) we should probably use that for further work. |
FreeBSD is getting hit by this as well: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=234828 |
Fedora is being hit too. |
All openSUSE releases now have this problem. If #4317 fixed the remaining problems with post- Heck, I'd be happy to try writing patches to fix any remaining issues, if it's necessary. Otherwise we won't be able to provide |
Fixed in #5636 |
As a quick fix to #4001 / prometheus/client_python#317, we pinned our prometheus client to < 0.4.0.
This introduces problems:
*_total
metrics.I'm still surprised at the prometheus_client devs making the choice to rename these metrics, but I think we're going to have to live with it, so could do with a migration plan.
When we have renamed metrics in the past, we have supported both old and new names for a few releases. That gives people a chance to update their dashboards and alerts. I believe we should do something similar here.
It may turn out that supporting both names is more easily done with prometheus_client 0.3 than it is with 0.4, so we should do this work before we are forced to update to 0.4 in a hurry which would break everyone's metrics.
The text was updated successfully, but these errors were encountered: