Jens

Hi Sorry for the late answer. I did try to tackle the whole problem at once. In retrospective that is most probably a step too big. I should have gone...

> > add jdk http client > > This will be a really nice addition. Especially if it's easy to configure an NioEventLoopGroup as the executor I'll keep that in...

Part of #251

@cjbooms this is ready for merge with the caveat that I would consider it experimental: things will be missing. I've tried to reuse the okio test cases.

It seems to happen most frequently when a run is stopped. I will comb more thorough the logs later

I went through the logs. To give particular example: I can see the worker coming online. Master node then logs out that the worker is available. Approximately 3 microseconds after...

Is there a specific class I could set logging to tracing on or something similar that would allow a bit more insights?