Retain old JobResults

[kkthxbye-code]

NetBox version

v.3.1.9

Feature type

Change to existing functionality

Proposed functionality

Currently all old JobResults are deleted when a new JobResult is in a terminal state. I propose that old JobResults are preserved.

Use case

Deleting old JobResults make it hard to use for automation through the API. Imagine a case where two runs of the same script is triggered via. the API. The API returns an id for the JobResult that can be used to query for the result. The RQ worker will run the scripts one after the other. When the first script finishes, you have the runtime of the second script to get the result. If you do not retrieve the result before the second script finishes, the JobResult is deleted and impossible to retrieve.

Also contributing to the issue is that JobResults are not webhook enabled, which could provide a workaround in some cases. But that's a different feature request.

Database changes

None

External dependencies

No response

[kkthxbye-code]

I have a working implementation shortly as a PR.

[jeremystretch]

I think this makes sense. The first question that comes to mind is: How will we handle the eventual cleanup of old JobResult instances? Presumably via the housekeeping management command? We'd also need to introduce a configuration parameter to dictate the expiration age similar to CHANGELOG_RETETION.

Also contributing to the issue is that JobResults are not webhook enabled, which could provide a workaround in some cases. But that's a different feature request.

#8958

[kkthxbye-code]

I actually did consider adding it to the housekeeping command, but figured that as JobResults, opposed to changelog entries, are deletable in the admin panel, it wouldn't be nessecary.

I can add it if you think it would be the best way to handle it. How about default retention? Unlimited would be my preference , but I'm not sure how much data people are actually outputting from their scripts.

[jeremystretch]

IMO unlimited (None) is a reasonable and safe default.

[kkthxbye-code]

I added a JOBRESULT_RETENTION option, added the functionality to the housekeeping script and updated the documentation. Let me know if I missed anything.

[jeremystretch]

Since this is a fairly substantial change to the existing behavior, I'd like to tag this for v3.2 (due to be released the week of April 4th).

@kkthxbye-code would you mind rebasing your PR off the feature branch? I don't think you'll run into much conflict, but please let me know if you need any help.

[kkthxbye-code]

@jeremystretch - I haven't rebased to a different parent before, could you maybe give a little hint. My google-fu comes short here, as there seems to be a bunch of different ways to get the same result.

I've tried git rebase --onto origin/feature origin/develop and then git add netbox/extras/management/commands/runscript.py as that file was shown as "deleted by us:" for some reason. Then force pushing with git push origin HEAD:save-job-results --force. It looked kinda correct, but when I viewed the pull request on github it indicated merge conflicts and it was still trying to merge into develop.

I'm kinda lost I guess :)

[DanSheps]

Should we target this for v3.3 now or release it as a minor in v3.2?

[jeremystretch]

I think it's reasonable to implement this in v3.2, given that:

• There are no changes to the reports or scripts APIs
• The change will be largely transparent to users
• I can't think of any way this would disrupt existing workflows

However, perhaps we should alter the default JOBRESULT_RETENTION value from 0 (indefinite) to 90 days (matching the default for CHANGELOG_RETENTION). This will ensure that even an oblivious user doesn't end up with an enormous backlog of job results, provided they are running the housekeeping daemon as advised in the documentation.