Improve Error Messages?

[klannon]

As reported by @ywan2:
Error messages like this:

accumulating task id 25313 item accum_25312 with 944356 events on [d12chas569.crc.nd.edu](http://d12chas569.crc.nd.edu/). return code 0 (monitor error)
allocated cores: 2.0, memory: 10800 MB, disk 14400 MB, gpus: 0.0
measured cores: -1.0, memory: -1 MB, disk -1 MB, gpus: -1.0, runtime 204.0 s
WARNING: task id 25313 item accum_25312 failed: monitor error
accum_25312 without result. accumulating: [('UL16APV_WWTo2L2Nu_NDSkim',
'file:///project01/ndcms//store/user/awightma/skims/mc/new-lepMVA-v2/central_bkgd_p1/2016APV/v1/UL16APV_WWTo2L2Nu/output_540.root'
, 'Events', 76218, 101624) p_25095 with result.

Are difficult to interpret. As I understand, a message like this can be generated because the monitor fails to receive data, possibly because disk space was exhausted or some other problem that causes the monitor to fail to extract monitoring data (see the -1 values in the message). It would be better if the message clearly identified that the error is related to the inability of the monitor to extract the necessary data.

Note: I believe (@ywan2 can confirm) this error message was generated with coffea 0.7 + WQ, so if that makes it obsolete because it's fixed in the new coffea + DaskVine code, then you can close this issue without objection from me.

[dthain]

@btovar what do you think?