All, There's been a couple cases of ABORTED builds due to timeout (e.g. [1])
In this specific case, the build was aborted after 16 minutes. The job *was* configured to use the "elastic" build timeout, which learns the expected build time based on the last three successful runs and allows an extra 50% margin. Based on this, it *should* have had a timeout of approximately 25 minutes, so I don't know why it aborted at 16 minutes. But to reduce this type of false negative, I've reconfigured the jobs to have an absolute timeout of 40 minutes. Builds are currently taking 15-25 minutes to run so this gives us a reasonable but not excessive headroom. Richard. [1]https://builds.apache.org/job/incubator-brooklyn-pull-requests/625/
