Jersey problem - possibly about systemd limits

hi,

I stumbled into this error:
2017-03-09T09:01:09.875+02:00 ERROR [ServerRuntime$Responder] An I/O error has occurred while writing a response message entity to the container output stream.
org.glassfish.jersey.server.internal.process.MappableException: java.io.IOException: Connection is closed

During this, the UI was sluggish, but the server showed no real resource constraint (mostly idle). According to this:
http://stackoverflow.com/questions/22962774/detecting-client-aborting-request-jersey

the problem could be the limits. So I changed the limits in /usr/lib/systemd/system/graylog-server.service
to
LimitNOFILE=102400
LimitNPROC=102400

Let’s see if that is gone. If this is indeed helped.

It seems there is a closed Github issue also about something similar:
https://github.com/Graylog2/graylog2-server/issues/2345