The worker is started with a millisecond timeout, with the assumption that it is up before the connection attempt. Sometimes it is not.
Simple to reproduce consistently - just wind out the timeout and it fails every time with the message observed in the CI failure.
The worker is started with a millisecond timeout, with the assumption that it is up before the connection attempt. Sometimes it is not.
Simple to reproduce consistently - just wind out the timeout and it fails every time with the message observed in the CI failure.