This morning the majority of the production jenkins agent were migrated from aws/eks to running on the ldf commons stable k8s cluster. Testing has shown that some parts of a build may run faster or slower in this environment. As an example, afw seems to be build ~2x faster due to the availability of more cores over the modestly sized aws instances that were being used. While operations that are heavily I/O bound, such as building a docker image, are expected to run ~2-4x slower.
If you notice any unusual failure modes or feel that a build is taking an excessive amount of time, please report it on #dm-square or via jira.