Thanks for your report, sorry about this inconvenience.
The issue is known, but it’s not entirely the step, it’s the backend server managing it. The issues are intermittent, rebuilding should solve this basically all the time, but the errors are getting more frequent we’ve noticed.
We are currently discussing how could we eliminate this happening completely but it requires a lot of planning. In case you are experiencing this frequently a day and in consecutive builds (so a rebuild didn’t help) please contact us on Support directly so we can look into rectifying that.