Sys-Admin: Daniel Appel
Fixes don't normally get pushed out straight away. Instead, they go into the next release (where they can have thorough testing & be rolled out without extra work).
If you want a fix to be pushed out faster, let people know by asking for a #hotfix. If the dev-team think a fix needs to be pushed out, they will let support (AMs + dev) know by emailing that we plan to do a hotfix.
Hotfixes get backported into the stable code (and hence every subsequent server update).
But we will normally only update the server for the client affected.
There's always some risk of a patch introducing new bugs, and limiting the servers that get upgraded reduces the possible impact.
Every so often, we will do a cluster-wide update -- either for a more serious or widespread issue, or just to bring everyone up to date after a series of minor fixes. The sys-admin will let AMs know when they happen.
By default, server updates will happen outside business and peak hours. This minimises disruption. Sunday 3am is the preferred time, so that any issues can be picked up first thing on Monday.
If the issue is urgent, we will of course do updates before then.
When an update is scheduled -- the sys-admin will send a short simple email to say what's being fixed, which servers will be updated, and when.