IP Filter: recode and put on Forge (was: after a reload plugins stops working)

After a reload in which nothing seems to have changed judging by the json in the audit log, the filter stops working. Further reloads or even restarts don't seem to fix this state.

One time this happened, it started working again a couple of days later. No restarts, deploys, reloads etc. had taken place around that time. In both cases this was cluster-wide, not node specific.

When setting the filter to debug, the following is logged for each request:

19.04.2017 10:49:23 [catalina-exec-20] DEBUG [org.onehippo.forge.ipfilter.IpFilter.requestData():422] Last event sent sent: 3004 ms ago, skipping this event
19.04.2017 10:49:23 [catalina-exec-20] DEBUG [org.onehippo.forge.ipfilter.IpFilter.allowed():149] No configuration match for host: <the hostname we used>

The hostname does match one of the regexes configured in the filter. Adding it explicitly didn't get rid of this message either.

After a while the first line is replaced by the following:

19.04.2017 10:49:47 [catalina-exec-24] DEBUG [org.onehippo.forge.ipfilter.IpFilter.requestData():411] Reached max number of event data requests 25