[Wikitech-l] Fwd: What to know before next week's shortening of the deploy cadence

Date: Tue, 2 Jun 2015 13:44:54 -0700 From: Greg Grossmeier &lt;greg(a)wikimedia.org&gt; To: Development and Operations engineers &lt;engineering(a)lists.wikimedia.org&gt; Subject: What to know before next week's shortening of the deploy cadence All: Next week we start the shortened deploy schedule. This means that code you write and merge into master will get to all of our users quicker than it has before. There are a few things I need to share with everyone so that we're all on the same page about expectations. 1. The first week or so might be bumpy as we adjust to the new cadence. Let's not call it a failure if things go bad next week; I fully expect some stabilizing time. I will be assessing things as we go and will not be afraid of reverting our process back if things don't go smoothly. 2. We (Release Engineering, namely Mukunda) will be more diligent and OK with reverting/holding the train if things start to look bad. We're going to be paying close attention to the fatal monitor in Logstash[0], and if there is a spike post-deploy that we can't identify and fix immediately, we reserve the right to quickly revert and then assign a task to your team to fix. <https://logstash.wikimedia.org/#/dashboard/elasticsearch/fatalmonitor> 2a. This is a great time to clean up your team's/projects log errors: <https://phabricator.wikimedia.org/tag/wikimedia-log-errors/> 3. We will continue to improve our automated and manual testing practices. Your help here is appreciated as at the end of the day you're the ones who know your code best. 4. Lastly, I think this exchange between Jon and I really encapsulates what we're doing here: <quote name="Jon Robson" date="2015-06-02" time="10:58:13 -0700"> Here we go :) Greg -- | Greg Grossmeier GPG: B2FA 27B1 F7EB D327 6B8E | | identi.ca: @greg A18D 1138 8E47 FAC8 1C7D |