I would like to analyze two different sources to determine how much data is being indexed.
index="_internal" source="*license_usage.log" s=*win*| stats count sum(b) by s | rename "sum(b)" as total_bytes | stats sum(total_bytes) | rename "sum(total_bytes)" as total | eval gb=total/1024/1024/1024
For the search above, there are 5 or 6 sources, so I want a count of the total gb used and this works.
What I really want is to be able to change the search to something like this...
index="_internal" source="*license_usage.log" s=*win* s=abc| stats count sum(b) by s | rename "sum(b)" as total_bytes | stats sum(total_bytes) | rename "sum(total_bytes)" as total | eval gb=total/1024/1024/1024
BUT, I want a sum of the s=*win*
and a sum of s=abc
Ideas?
like this ?
index="_internal" source="*license_usage.log" (s=*win* OR s=abc)
| eval s=if(s=="abc","abc","win")
| stats sum(b) as gb by s
| eval gb=gb/1024/1024/1024
what systems ? where does 31 come from ? how will "per s" work given we're wilcarding win to get the sum ?
You need to be more specific with your examples
ok - so to add to complexity....
This work great, I would like to take my count of systems per s and determine the mb used per system.
I was using something like to but I had to run a search for each group.
index="_internal" source="*license_usage.log" s=abc| stats sum(b) by s | rename "sum(b)" as total_bytes | stats sum(total_bytes) | rename "sum(total_bytes)" as total | eval gb=total/1024/1024/1024 | eval mb=total/1024/1024 | eval mb_per_system=mb/31