Closed h-vetinari closed 1 year ago
We've been having other issues today ( https://github.com/conda-forge/status/issues/137 ). Would not be surprised if the churn from that (either through CI builds or token rotation) is causing Travis CI issues
Travis is really struggling at our scale. I think we're gonna have to rethink how we use Travis and how we provision feedstocks on it.
How are we doing with this?
A couple of jobs on feedstocks I maintain were created today for the aarch/ppc migration, and builds like this one have been waiting for almost 3h for a place in the queue, but nothing seems to start. According to the status page, travis only ran 1 job in the last 8 hours, so it does not sound like an oversubscription issue.
I was wondering if that's a known issue?