We have discussed in the past how to tackle changes in Opencasts endpoint configuration, f.e. disappearing of services and being taken over by other machines. One idea was, to rediscover endpoints on a regular basis, so i added them to two cronjobs which are run regularly.

This now seems to introduce other problems, like the ones described in this issue, and makes even more problems than the other approach, which is only updating endpoints if someone adds or edits an opencast server, like in V2 of the plugin.
The downside to that is, that it might be necessary to go into the settings and do exactly that after an plugin update.

It seems, that the original approach is still better then trying to rediscover the endpoint every 10 minutes. My suggested solution to the problem would be for now to remove the auto-update of the endpoints and live with the nuisance rather than the nasty errors.