Conditionally disable access to update manager routes

This is effectively the same as removing the setCacheMaxAge(0) call in UpdateManagerAccessCheck::access. Should we just remove that instead?

One comment on the MR but looks great otherwise.

Deprecated the access check and added a CR. Not sure if versions are correct.

Hey sorry but can the issue summary be updated also.

Just reading the summary I wouldn't of known we are deprecating UpdateManagerAccessCheck for example

Updated the issue summary.

Thanks! CR also reads well.

All questions here are answered, I read the MR and CR and they are clear. The only thing left would be changes for a 10.4 backport, but that doesn't need to be done now.

catch → closed merge request !8620

Committed/pushed to 11.x, thanks!

I was about to ask for a backport MR, then I realised it would just be the same change with the first hunk of the diff removed, so went ahead and made that change locally for both 11.0.x and 10.4.x. Reclassifying as a bug report since this fixes performance issues.

Any chance of a back port to 10.3 where the regression was introduced?

Was a bit hesitant because of the new class, but I think it's fine and better than shipping the bug for 11 months - backported to 10.3.x too.

Automatically closed - issue fixed for 2 weeks with no activity.