Increasing that limit will probably have adverse impacts on performance. If there is a problem with condition limit, you should see whether there is room for improvement in the filters before changing this.

Also are you actually sure that there are changes which should have matched these filters before considering that it's a technical problem?

See https://www.mediawiki.org/wiki/Extension:AbuseFilter/Conditions for more background.

What we could do is to increase the limit 24 hours, and measure impact.

If there is a problem with condition limit, you should see whether there is room for improvement in the filters before changing this.

Already done (long before I started this request).

Also are you actually sure that there are changes which should have matched these filters [...]

yes, several times during the last year. Although I don't know, whether the condition limit is reached, because I can't see any hint on that. (Formerly there was a message per rule saying how many conditions in average it took. But this messages seemed to be like random number and I guess that's the reason why they have vanished.)

Maybe better than trying out something for a few hours would be some measuring of the reached conditions over time. Is that possible? Is there a log of number of conditions per time/edit?

@Gilles would a 24 hours test per T129171#2098988 be acceptable?

I think @aaron will have a more formed opinion than me on the matter, I've never been around that code.

This might no longer be necessary after my recent fixes to the way conditions are counted (which generally decreased the number of conditions used by most filters), see longer explanation at T132048#2212440. https://de.wikipedia.org/wiki/Spezial:Missbrauchsfilter is currently saying that 0 recent actions exceeded the limit (but this doesn't tell is much due to T132925).

It's also always possible that there are mistakes in the filters which cause them not to match. ;)

In T129171#2101016, @seth wrote:

(Formerly there was a message per rule saying how many conditions in average it took. But this messages seemed to be like random number and I guess that's the reason why they have vanished.)

This can be re-enabled now on a per-wiki basis, at least temporarily (I'm not sure how big is the performance impact of recording this data on every action, but it's probably better not to leave it on forever). See a similar change for Commons in T132200.

Another option is to enable $wgAbuseFilterProfile for dewiki to help them identify which filters are so condition-demanding and have them tweaked. I was adviced in a recent request to raise the condition limit of Meta-Wiki to try that first.

If dewiki thinks there is a problem, I support profiling enabling first to debug it and if all filters are needed, then this variable can be raised.

@seth Can you please confirm that this is still a problem? Thanks.

Hi!
Sorry, but I can't confirm, because during the last months I used the edit filter very(!) rarely. (So I can't confirm the opposite, too.)