Comment by InternalRun

I’ve experienced the same as you, I’ve lost many hours to fixing different protocol time out errors. We’re running it in ECS now and it’s pretty stable, every now and then one pops up but is fixed by retrying the job.