Summary: "Max number of days to keep history entries" isn't honored

The JobConfigHistory plugin records any remote agents (slave nodes) that are added to, modified or removed from Jenkins, which is all good ... but, as mentioned in JENKINS-62281, the exclusion filter isn't applied to node records - the user can't control whether or not history is recorded.

However, the plugin then fails to honor the "Max number of days to keep history entries" field, and that can then fill up the disk, resulting in JENKINS-35421 (and worse).
JENKINS-61199 reports that the plugin doesn't honor this for deleted jobs either, suggesting that the problem may not be just node records - it may be more of a general issue.
i.e. it may be that once something is "deleted", it's forgotten entirely and will remain on disk forever, regardless of the "Max number of days to keep history entries" setting.

...and that then means the plugin can unwittingly fill up the disk and "break the world".

This is a minor problem if nodes (or jobs) are only ever created & deleted infrequently, but when things are more automated, this "theoretical" problem rapidly because a show-stopping major problem capable of crashing Jenkins if one does not manually clear out the config history folder.
That kind of clear-out shouldn't be a manual task - that's the sort of thing the plugin ought to do automatically - it needs some form of background cleanup process.