Description

We have a server running and were experiencing some issues so decided to use spark, which is why we setup the server to automatically run /spark heapsummary every 30 minutes to check how it performs overtime and compare results that way, however, we noticed that after adding this, for some reason all players started to have 1000 ping and getting kicked for timeout. This was never an issue for us and it is not the hosting provider we use as it only started happening after enabling that, and it fixes itself when removing that.

Reproduction Steps

• Install spark
• Make it execute /spark heapsummary every 15-30 minutes
• Wait

Expected Behaviour

It should not affect player connectivity at all

Platform Information

• Minecraft Version: 1.20.4
• Platform Type: Server under a proxy
• Platform Brand: Paper
• Platform Version: 9e171ef

Spark Version

v1.10.34

Logs and Configs

No response

Extra Details

No response

(It is not always at 1000ms, it happens every 1-2 hours in our case with it running).

Java 17

You could try profiling all threads (/spark profiler start --thread *) during the execution of the command (start shortly before, stop shortly after). Maybe that provides any useful information what's going on.

That command will typically trigger a GC run, and the duration of collecting the required data also depends on the (live) heap size. It shouldn't cause people to time out.

What Java version are you using exactly?