[17:00:31] <greg-g>	 Hello and welcome to the deployment office hours. Please feel free to ask your deployment during COVID related questions here
[17:01:03] <thcipriani>	 related guidance outlined in https://wikitech.wikimedia.org/wiki/Deployments/Covid-19
[17:10:03] <bd808>	 o/
[17:10:33] <thcipriani>	 bd808: are you waving or in the queue?
[17:10:58] <bd808>	 just waving, but I think I have a question too
[17:11:35] <thcipriani>	 sure, shoot :)
[17:11:44] <bd808>	 My question is if these policies are mainly directed at MediaWiki components involved in the Wikimedia production deployments?
[17:12:23] <bd808>	 I'm thinking about all the things my team does outside that envelope and wondering if we are expected not to ship changes to Toolforge, etc
[17:13:36] <thcipriani>	 this guidance was mainly formulated by thinking about WMF production deployments, yes.
[17:14:36] <thcipriani>	 I think once key question that is generally applicable is the question of scope: being mindful that folks affected by your deployments may be less available in the near future.
[17:16:19] <greg-g>	 yeah, that ^
[17:16:51] <bd808>	 I don't want to open up a big can of worms, but is there a guiding principal at work here from the RelEng side about risk management? By that I mean that broadly there are 2 camps: ship in small increments often, and ship in well tested increments less often. This seems to be signaling the latter, and in alignment with things like the typical December freeze.
[17:17:19] <bd808>	 I'm wondering if there is an existing essay about how RelEng looks at this problem/tension
[17:17:45] <greg-g>	 We want smaller things more quickly, which is why we're not stopping deployments
[17:18:12] <thcipriani>	 there is an inherent tension here, though
[17:18:43] <bd808>	 I 100% agree that it is a situational and nuanced issue
[17:21:50] <thcipriani>	 indeed. As a direct answer there is no essay about our thinking. We do, broadly, want smaller things to go out so that deployments are easier to reason about. SAtopping deployment is definitely something we'd like to avoid.
[17:22:02] <thcipriani>	 *stopping
[17:27:33] <RhinosF1>	 Based on that page, easily testable swat-able config changes can go ahead as normal right?
[17:27:55] <thcipriani>	 RhinosF1: yep, that is correct
[17:28:27] <RhinosF1>	 Cool, guess I’m probably swatting at some point then
[17:28:42] <thcipriani>	 great :)
[17:29:02] <RhinosF1>	 going to look at https://phabricator.wikimedia.org/T248311 tonight