This might also help.
http://www.splunkbase.com/apps/All/4.x/App/app:Splunk+License+Usage
You can tweak the xml to your needs.
If syslog is one of the top ten sourcetypes or sources, then you can use the metrics log search below:
index=_internal source=*metrics.log* per_sourcetype_thruput series=syslog | eval gb=kb/1024/1024 | stats sum(gb) as Total
If you want to see a charted breakdown, you can substitute the timechart command for stats:
... | timechart span=1d sum(gb) as Total | addtotals
For the above searches, use the TimeRange picker to select the 30, 60, or 90 day duration. If syslog is not in the top ten sourcetypes, you'll have to do a somewhat expensive query that measures the raw size of the events. So something like:
sourcetype=syslog | eval size=len(_raw) | stats sum(size)
Something like this will give you what you need.
index=_internal (source=*/metrics.log* OR source=*\\metrics.log*) group=per_sourcetype_thruput series=syslog | timechart sum(kb) by series
Then, just change your timeframe to 30, 60 or 90 days appropriately.
Sadly, I believe the default setting for _internal is 28 days or 2419200 seconds. You can raise that setting 'frozenTimePeriodInSecs' to whatever fits for you so that you can get a greater field of view.
If there is no coldToFrozenScript set, it simply erases the data.
BTW - Thank you for your help with this. It's greatly appreciated.
The above query works; however anything after 30 days there is no data showing up. This system has been in production for several months. Could it be with logs being rotated? If so, how to proceed?