2 hours ago · Tech · 0 comments

We’ve been working to a pretty tight deadline as a team recently. We had a period of relative stability with a few concurrent projects on the go, and then we had this new elephant-sized project dropped on us. The entire team has rallied around it and everyone’s contributing exceedingly well to their own parts. During the last few months of this project, I’ve been taking a look at how we can get to the point of verifying our changes faster after reading Accelerate for the second time. Trying to find answers for how we from “push” to knowing something works, and then shipping that with less time passing between all of those points. A few things that our team has worked on to fix this has been: Improving Docker build caching so that our tests start running sooner (8min -> 2min startup) Addressing flaky specs as they’re detected, rather than backlogging them as a fix for “later”. Specs are split into groups, and each group would take 10 mins per retry, retrying up to 3 times. Adding a…

No comments yet. Log in to reply on the Fediverse. Comments will appear here.