Some of our Deployment Servers going offline with these events flooding SplunkD.log:
WARN HttpListener - Can't handle request for /services/broker/connect/<GUID>/<Server>/271043/windows-x64/<port>, max thread limit for REST HTTP server is 2728, threads already in use is 2728
The servers eventually show as Offline and when they're rebooted they resume working for a while.
Why is the limit so low? I found in the documentation where this is addressed in 6.3 (http://docs.splunk.com/Documentation/Splunk/6.3.0/Troubleshooting/HTTPthreadlimitissues) but can't find where to address this in 6.2.6, and we're not quite ready to upgrade.
The workaround should be the same for 6.2.6 as 6.3.0. The thread limit was introduced in 6,0.0 so the calculation and override are the same for all 6.x.x versions to the present.
View solution in original post
I have same issue except I have Splunk Version 6.3.0 loaded on windows. The workaround is for Linux users. Is there a workaround for windows I wonder.
Thank you, I've set it to -1 on the server that was having issues and since haven't seen any errors.