currently the slaves graph shows failures, retries, total (failures+retries). I would like to include passes, and infra as well, then the total would include all of the jobs types.

I might got it all wrong, but that kind of report will not be readable. Selecting distinct slave names from ouija database where result is "success" and testtype is not a blank string returns more than 1000 rows. Perhaps, that should be done with slave type (as in bug 909798)?

This should be a single value, not an entire list. The ultimate goal here is to figure out which slaves are failing an abnormal amount so we can pull them from production and reimage them. While we can tell that based on total failures, it also helps to know if we end up with a small number of machines running thousands of jobs with <1% failure rate even though those total number of failures might look high compared to other slaves running much fewer jobs.

(In reply to Joel Maher (:jmaher) from comment #2) > This should be a single value, not an entire list. The ultimate goal here > is to figure out which slaves are failing an abnormal amount so we can pull > them from production and reimage them. While we can tell that based on > total failures, it also helps to know if we end up with a small number of > machines running thousands of jobs with <1% failure rate even though those > total number of failures might look high compared to other slaves running > much fewer jobs. So, for every slave listed on slaves page, I need to add additional column that contains "Passes" (this should be the first column after slave name?) and include value of this column for total runs calculation? I don't need to show slaves that have only "success" results?

It should be the last column in the list of non calculated data, as it is useful but not the most critical. If a slave has no failures, it would be nice to show them as well, a big bonus would be a checkbox that could hide/show non failing slaves.

Comment on attachment 806723 [details] [diff] [review] 0011-add-new-column.patch Review of attachment 806723 [details] [diff] [review]: ----------------------------------------------------------------- Looks good, thanks!