Hi there,
We would like to track some additional troubleshooting information in the tasks’ data, that would help us identify issues in simulations etc. That would save us a lot of time from having to go and parse logs for that info, as we could have an overview of the most important numbers from the monitor directly.
cheers
laszlo
Hey Laszlo,
We actually have this on the wish list already, but one issue we’ve struggled with solving is how to display this information in the Monitor. Odds are, you’d have different extra info values for different job types (or even within the same job type), so having Extra Info columns with customizable names becomes problematic because the names would almost have to be customizable at the job level (not the global level like it is for Job and Slave extra infos).
However, it’s not an issue if we were just storing arbitrary extra info dictionaries that are never visible in the Monitor lists (like we also do for jobs). Is that what you’re looking for, or are you actually looking for a way to view this information in the Monitor. In the latter case, any thoughts on how you imagine the data would be displayed?
Cheers,
Ryan
Hi Ryan,
Yes, ideally we would have the information visible in the monitor, so you can compare different tasks’ easily.
To us, customizability of the column names is secondary, so i wouldn’t mind if they are simply named as Ex0, Ex1, Ex2 etc., without the ability to change that name per job.
Playing with theoreticals though - in a perfect world - the job info file could contain the names you would want for the columns, should you want to override the default names. Something along the lines of:
TaskExtraInfo0Name=Element count
TaskExtraInfo1Name=Bucket count
etc.
If they are not overwritten by the job, the original names would be displayed. The ‘view’ in control of the task panel could override the column headers, should it need to.
But as i said, we would be perfectly happy with non-customizable column names as well, as long as we can show them and populate the fields with our data.
Hey Laszlo,
That’s actually along the lines of what we were thinking. It feels a little clunky, but maybe that’s okay, and you’re right that we could always just fall back to Extra Info 0, Extra Info 1, etc.
I think we can put this on the tentative roadmap for the 8.x dev cycle. Not sure if it could fit into 8.0, but it might work as an 8.1 feature.
Cheers,
Ryan
Awesome, thanks Ryan!
+1 This would be v useful.
For now I’m thinking of storing extra task data in a Job ExtraInfoKeyValue field using a JSON string.
Make sure you watch for our press releases around Siggraph… 
Yep, JobExtraInfo KVP could maybe do just for now.
cool