Have an environment where a directory is used to 'stage' files waiting for an update. Essentially, a file is sent to a vendor and a local status file (of zero or more bytes) is placed in a 'pending' directory. Once the vendor has completed their work and a good return is received the local status file is moved to the 'done' directory. (This is done so a second process can use the listing to determine what status file to check for on it's run).
What I want to do is monitor the '/pending' directory for new files only being worried about WHEN they showed up. Then, if they haven't 'moved' in a predetermined amount of time (depending upon environment this could be two - eight hours) raise an alert. This would also allow me to say "_time seen" and "_time gone" to build some metrics on vendor performance.
All that being said, I am having issues getting it set up to forward from my linux universal forwarder. My inputs.conf stanza looks like:
[fschange:/path/to/dir/]
signedaudit = false
index=_audit
I read that in one of the previous 'answers' questions. Unfortunately, I am seeing no records in _audit or _internal for that host using the search head.
Ideas?
You should not be putting data in _audit
, that is an internal index for Splunk to index its own events. I know that there is some stuff in the documentation that talks about this index, but unless you have already done all the other auditing setup, it isn't going to work.
Create a new index - you could call it audit
and send data to that index instead.
[fschange:/path/to/dir/]
index=audit
Also, this feature is deprecated: Monitor changes to your filesystem
You can still use it, but it might go away.