What is the impact of archiving records on PA scores and PA snapshots?

[Pierre P_ Mouri]

I have been asked to evaluate the impact of archiving records on PA scores and PA snapshots, and how it affects Analytics Hub and Dashboard viewers.

Here is a scenario:

• There are 15 new incidents created on May 1, 2020
• The daily job containing the Number of New Incidents indicator runs the next day, with "collect records" enabled. It captures a score (count) of 15 in the pa_scores table(s) and stores 15 sys_ids in the pa_snapshots table (we will ignore breakdowns for simplicity).
• We can see both the score and the records in the Analytics Hub
• X months/years later there is an archiving initiative and all the incidents created that day are archived.

My understanding of the situation post-archiving (based on Dev instance testing):

• The score for that day is still 15 (ie. scores are unaffected by the archiving of underlying records)
• The sys_ids are still stored in the pa_snapshots table (ie. collected sys_ids are unaffected by the archiving of underlying records)
• We can see the score (unchanged at 15) in the Analytics Hub, BUT the records are no longer visible in the Analytics Hub and for PA users in general.
• Running a historical job that covers archived periods would risk overwriting good data with bad.

Does this accurately reflect the impact of archiving record on PA scores and PA snapshots? Any missing pieces or blind spots?

Thank you!

[Adam Stout · Accepted solution]

I think everything you said is correct, but remember, daily snapshots are generally only kept for 6 months anyway (as opposed to the 2 years that scores are kept). This can be changed globally or per indicator, but those are the defaults.

Are you archiving before the 6-month window?

[Pierre P_ Mouri]

Thank you Adam! The use-case would be to run built a historical time series prior to archiving old (1 year+ I believe) records, and to keep building (via daily collection) the time series in the context of regular archiving going forward.

I understand that I will have to update the default setting for scores retention if we are to build a time-series that is longer than 6 months 2 years. Keeping snapshots for records that are archived appears of little value, except maybe from an audit perspective.

[Adam Stout]

The daily scores are kept for 2 years by default, not 6 months.