Ingredients of a Scalable Merge Queue | Merge Queue 101
If you're not using a merge queue, you merge untested code daily. If you're using an underfeatured merge queue, you're going to hit a wall where your CI costs are crazy high, and it takes hours or days to merge things.
Here're features your merge queue needs to scale and keep up with your fast moving team.
0:00 Introduction
0:36 The challenges of scaling a merge queue
1:20 The basics of a merge queue
4:04 Predictive testing (pre and post-commit tests)
4:48 Optimistic merging
5:23 Dealing with Flakey Tests
6:12 Batching PRs to save CI time and cost
7:17 Parallel queues for independent changes with Bazel and NX
8:00 Large mono repos with Trunk
8:18 Conclusions
Find us on Twitter/X: https://x.com/trunkio
Find us on LinkedIn: / trunkio
Find us on Slack: https://slack.trunk.io/
Learn more about Merge: https://trunk.io/merge
This video dives into the challenges that large engineering teams face with merge queues during code integration, such as long queue times, test flakiness, and inefficient handling of minor changes like documentation updates. Introducing Trunk, a tool designed to improve merge queues by enabling predictive testing, concurrent testing, optimistic merging, and batching, which collectively reduce integration issues and save time. Essential for scaling in large monorepos, Trunk's capabilities are illustrated through examples, emphasizing the tool's common-sense approach and efficiency. For more details or to book a demo, head to https://trunk.io/merge.