Highlights

• 🚀  Flaky test management - beta access

• 🚀  Introducing Aviator Pilot

• 🚀  Slash commands to specify affected targets

• 🚀  Configuration audit log

New features

Flaky test management - beta access

What it does:

• Connect with your CI to automatically detect most disruptive flaky tests

• Configure a status check on GitHub that suppresses flaky tests. This works with Aviator merge queue to reduce your rerun burden

• Run automatic nightly builds to track new flaky tests

Learn more here. If you are using CircleCI or Buildkite and want to try out FlakyBot, reply back.

Introducing Aviator Pilot

We understand that every team has a unique way of managing dev workflows. So we are building a configurable automation framework called Pilot! Things you can do with Pilot:

• Build your own rules for multi-queue dispatch, emergency fixes, priority queues

• Take custom actions on forked PRs or based on author’s GitHub teams’ membership

• Send Slack notifications to specific users on custom events

Interested in early access to try out Pilot, reply back.

Affected targets slash commands

Now you can specify affected targets using slash commands. This is very helpful if you want to define introduce multi-queues within your workflow. Learn more.

Configuration audit log

Now Aviator captures any changes to your Merge rules configuration via UI or GitHub config file. You can see changes to your Aviator configuration right from our Merge Rules page in the UI.

A couple of things to note here:

• Changes made by Github users will have the Github username and a link to the base branch commit that introduced the configuration change

• Changes made by Aviator users from the UI just store the aviator username (which typically will be an email)

Bug fixes and improvements

• Now you only need to apply a single label to merge with custom behaviors. For e.g., if you only apply a skip label or a rebase-override label, that will also notify Aviator to queue the PR while handling the custom behavior.

• To stay consistent with how Github treats statuses, we now treat skipped as a success status. If you want to override this and use skipped as a failure status, you can use our required_checks configuration. See docs.

• Fixed auto-requeue issues for PRs with merge conflicts

• Fixed rebase for forked repos

• Stacked PRs improvements