I have some summary index data that is stored with sistats:
index="_internal" group="per_host_thruput" source=*metrics.log | eval megs=kb/1024 | fields series,kb,megs | sistats sum(megs) by series
I'd like to do a delta on some of this data but I'm not quite sure which events/fields I should use since sistats does some massaging of the data.
index=summary prestats_reserved_type=sum | delta prestats_reserved_val as delta_megs
This will get the delta info that I need, but
index=summary prestats_reserved_type=sum | delta prestats_reserved_val as del | stats max(del) by series
Returns nothing. Any thoughts?
When using a summary index created using the "si*" sufficient statistic accelerators, the only supported operation on the data after retrieval is the corresponding full reporting command. That means, for sistats <aggregator>
, you can only run index=summary source=... | stats <aggregator>
. The same holds true for sitimechart
, sitop
and sichart
.
The "si*" command output is an internal opaque format that is interpreted in a special manner by commands within splunk.
What's the goal report that you want here?
Thanks for posting this - I was doing | sistats count by host, source, sourcetype, field1, field2 and able to find host, sourcetype and source as having the orig_ rename, but couldn't get the count (in the psrsvd_gc field) to play nicely. It constantly said it was an "internal variable". I'll just use index="summary" | stats count by orig_host, orig_source, orig_sourcetype, field1, field2. I'll repost this as a question too. http://answers.splunk.com/questions/6887/
I am trying to create a report that notifies me when a host starts sending an unusual amount of data by using the summary index data. I suppose I could just use the metrics data as is