Filter _internal events that get forwarded

[belka]

My company has a lot of small, independent offices that have a handful of machines in each locations. The problem is there are literally hundreds of locations, many of them rural with low bandwidth.

I'd like to filter what gets sent to index=_internal by the UniversalForwarder. I know that a Heavy forwarder can do that, but I've already provisioned thousands of Windows machines with Splunk and the _internal index is clogging some of the pipes. Any suggestion how to get just warnings and alerts for _internal instead of bandwidth-filling info events?

[kristian_kolb]

Perhaps this can be of help? Either stop forwarding the _internal events from the forwarders altogether with the forwardedindex setting in outputs.conf, or configure the logging levels for the various components through log.cfg;

http://docs.splunk.com/Documentation/Splunk/latest/Admin/outputsconf
http://docs.splunk.com/Documentation/Splunk/6.0.1/Troubleshooting/WhatSplunklogsaboutitself