cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Tracking Minor Versions (Application Installs) best practices

Does anyone have a best practice for reporting on Installations for minor versions of applications?

  • For example, tracking installs of Tableau Desktop 2021.2.8 vs 2021.2.17
  • Since Flexera rolls that up to the 2021.2 main version, the minor versions .8 & .17 don't get split out.

We could go into each evidence list one by one and track, but that is very cumbersome.

I see there is an ideas submission that suggests adding evidence to the report builder, but is not going to be added due to size of database (which makes sense).

Should I try:

1) adding a dummy app for just Tableau 2021.2.8 (and assigning that incremental evidence) - this is what Nicolas suggests in the Ideas. (However I am not sure if flexera's existing ARL supported 21.2.% wildcard evidence will mess with my results)

(assuming this might conflict with our existing installed apps data though)

2) is this something the "application transparency report" via the "detailed evidence" flag solves?

Thanks!

(1) Solution
nrousseau1
By Level 10 Champion
Level 10 Champion

Hello @johnksilverwood ,

Indeed this report provides the "full recognition story" but focuses only on the applications that have been flagged (to avoid the billion of possible rows in the ImportedInstalledFileEvidence table for instance) and also to remediate a possible synchronization issue: let's say SCCM was imported at 8:00 AM and the next full inventory import and reconcile happens at 10:00 PM. The imported raw evidence (particularly newly installed applications) may not match the installed applications that may have not yet normalized based on the latest SCCM import.

You can add the flag by batches from the "all applications" screen that has the "Import Detailed Evidence" button.

If you prefer something more straight forward, are on premise and accept the risk of data de-synchronization, you can use the report published the the Community.

I hope it helps,

Nicolas

View solution in original post

(4) Replies
nrousseau1
By Level 10 Champion
Level 10 Champion

Hello @johnksilverwood ,

Indeed this report provides the "full recognition story" but focuses only on the applications that have been flagged (to avoid the billion of possible rows in the ImportedInstalledFileEvidence table for instance) and also to remediate a possible synchronization issue: let's say SCCM was imported at 8:00 AM and the next full inventory import and reconcile happens at 10:00 PM. The imported raw evidence (particularly newly installed applications) may not match the installed applications that may have not yet normalized based on the latest SCCM import.

You can add the flag by batches from the "all applications" screen that has the "Import Detailed Evidence" button.

If you prefer something more straight forward, are on premise and accept the risk of data de-synchronization, you can use the report published the the Community.

I hope it helps,

Nicolas

And to comments on the "trick" I shared on the idea. Your tableau example will be addressed cleanly by the application transparency report: the Tableau 2021.2 recognized application will show each of the raw versions that have been inventoried.

My trick was more... you want to see the row versions of all SQLServer.exe files (while this one is not used by the recognition in ARL (that uses more WMI evidence). You could absolutely create a dummy SQLServerFromFile application, link it to SQLServer.exe, version "%" and you will gather all the raw data. To be used carefully as we don't want the billions of raw of a too loose file evidence definition.

Thanks @nrousseau1 ! We are cloud so I will focus on the Transparency report.

To your point about careful use of the flag (evidence tables are massive) - is there a best practice for limit on how many apps you flag? 

Assuming the answer is - "as few as possible" 😁 but just wanted to be sure.

I guess really the strategy is - only flag detailed evidence for a short period of time, and then go back and uncheck anything you think you dont need anymore.

I just wanted to make sure to look out for any signs of issues, but I guess as cloud customers, we probably won't see anything on the back-end. Maybe a slower reconcile or import process?

Thanks for your help Nicolas. 

John, we are Cloud too and have used this creation of a dummy application to be able to report on all related installer and/or file evidence down to the raw version for this exact purpose. Performance-wise, our current Transparency report has about 200-250 different applications switched on to collect detailed evidence, and our Transparency Report currently produces around 700,000 rows of results - and - the performance has been completely fine. Zero issues with jobs hanging, general performance hanging, or report run-times, export-times. All systems normal.

Best practice.

  1. Switch the evidence collection off after your use-case is closed. Just to be clean. Safe.
  2. This process is very likely to create Overlapping Application alerts. All the more reason to keep things tidy.