Hi all,
We're trying to configure winlog beats to drop info level logs but seem to be missing something. Ive found a post that is trying to do the same thing as we are but their config isnt working. So I'm guessing my logic is broken?
First, you need to check how is formated the name of the field for the log level in your data.
I have a doubt on log.level: 'information' , depending of the version of winlogbeat, it can be different.
Can you be more precise about what you want to achieve ?
You said you don't want all information level, but only for security Logs ? Or for every log categories?
Secondly you want to drop event id 5145 from security logs, right ?
So perhaps dropping information level, is less efficient (memory/cpu) for the agent, than configuring only the log levels you want before the drop processor.
If you want to have , for security logs everything except informational, and dropping 5146 whatever the log level :
So I checked on some logs, and the field log.level: 'information' is correct on the latest winlogbeat version.
Your config file should be working.
However you said it was not working as expected, how did you figure it ?
And what do you expect precisely ?
hmm, this sounds pretty strange actually.
Quick questions :
Do you restart the winlogbeat service after you modify the config file ?
Are you sure info logs are sent from the agent/machine you're trying to configure ?
so strange update. It works now. The weirdest part is all the changes I've made today are things we have done in the past. We've previously made the same/similar changes over a weekend and let it sit and nothing changes. But this time it worked. Must be user error on my part of a really strange bug.
Good news.
Make sure you always restart the service after any config changes.
As well the modification will only apply to new events ; "old" ones are already present in elasticsearch and changing config has no effect on these "past" events.
Sometimes we've had configs not take even after restarting the service which is strange. I have no idea.
Yeah I'm aware of that fact, normally we filter on last 5 minutes after waiting to account for old data still possibly showing up.
Thanks for the help and working with me. Its lovely having someone help as opposed to saying hey heres a link to the documentation that you've probably already read.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.