Is it possible to re-run historical PA data collection with a new filter?

[AngusWIlliams]

Hi,

I'm very new to working with Performance Analytics and was wondering if I would be able to recollect historic scores for an indicator in this circumstance:

We have a PA report which has been running daily for a few years and intends to look at the average time taken to respond to an Incident by way of its response SLAs. The indicator source is set up to find incident_sla records which have been opened on that day or before, and which don't have a resolved time (i.e. are currently open). The indicator then averages the business elapsed time of these records to get an hour score. 

The problem is that the way this is set up meant it didn't account for incidents which skipped resolved and were immediately closed (i.e. had a closed time), so those records (often with very large breached SLA business elapsed times) have been being repeatedly counted in the indicator and skewing the data to be far greater.

Is there a way for me to retroactively update all scores to exclude records which were closed at that moment in time from the calculation. Can re-running the historical job accomplish this?

Thanks for any advice in advance,
Angus

[Mark Manders · Accepted solution]

No. This will probably delete all of your data, because you run it and it will look at all incident_sla records that were opened on the day you run it and don't have a resolved time.

What you could try (do this on DEV first!) is to update all of those incidents with the closed time stamp on the resolved date (and make sure that's done from now on on any new incidents). Then run a 'repair sla' job on all of those records, updating those records. The next PA job run will not take those into account anymore. You can't correct the past, but your future will show correct data. 
I am just very curious what the business case is to know the average business elapsed time on open incidents.