Splunk Listen Backlog Queue

[chrisgoffient]

I have a client that is using Splunk enterprise using TCP, we've been monitoring the number of ListenOverflows, and increased net.core.somaxconn from the original 128. We noticed after a splunk restart, it looks like listen(...) is still setting 128. Is there a tunable in Splunk for TCP over 9997? I couldn't find it in the documentation. Most systems allow this to be tunable (i.e nginx, apache, etc).

[mwidjaja_splunk]

To make that change, you'll need to edit $SPLUNK_HOME/etc/splunk-launch.conf and add this line below
SPLUNK_LISTEN_BACKLOG=new setting

And restart