Connected nodes can be marked offline, such as in this case with the free disk space check failing. This is a feature. /computer/configure sets you disable that, or define custom thresholds.

Not enough information is provided regarding the builds; the screen shot does not include older builds. The message is used when the specific job does not allow building in parallel and a previous build is still running.

Please also disable the custom themes when providing screenshots in the future, it's unnecessarily difficult to understand what's going on.

danielbeck Thanks for looking into this.

I'm not sure what information you were missing about older builds. If you can elaborate on that I might be able to provide it. I didn't include all the builds because the build queue had gotten so large that the last built item was not visible on the screen.

Sorry about the custom theme. I'm not trying to be difficult, I just don't really have the option of disabling custom themes unless there's some way to do that only for one user? I have hundreds of users on the server, so changing the theme, even for just a couple minutes would confuse everyone. We have several instances running and we use the different themes to distinguish them.

I don't agree that a job displaying "build in progress" when no build is in progress is "not a defect". In this situation it was very difficult to figure out exactly what corrective action I needed to take to have the jobs continue building. 

I'm not sure what information you were missing about older builds. If you can elaborate on that I might be able to provide it. I didn't include all the builds because the build queue had gotten so large that the last built item was not visible on the screen.

Well, so far you don't show that there's no build running. There might be, just way down the list. A thread dump from master might be helpful here (I think what you provided is an agent?)

Or the output of /job/whatever/api/xml?tree=builds[number,building,result]

Ah, unfortunately the owner of the node has since cleared up the memory problem and restarted the slave agent. The reason I came to the conclusion that no build was running was from one of the smaller screenshots I posted. Usually it looks like this if a node is currently running something: 

I will keep an eye out for this problem in the future and get a thread dump from master next time. The thread dump I attached was from the slave agent.