- Mark as New
- Bookmark Message
- Subscribe to Message
- Mute Message
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
calculate time statistics over an hour, but only find releated events that occur within one minute
I am trying to calculate statistics for when a transaction enters our application, and when the reply is sent from the application. I would like to calculate statistics over an hour and there are two key values that i use to find the events to caculate on (research and Locsite). Here is the query:
sourcetype="Filter" transactionType=A44 | stats min(_time) AS earliest max(_time) AS latest by research,Locsite | eval duration=latest-earliest
this query returns the duration of the transaction. this query is good enough most of the time but sometimes events with the same "research" and "Locsite" is returned more than two times within the time range of an hour, then the duration value will be calculated over too long time. So i would like my query to only look for events with the same "Locsite" and "research" within one minute, but calculate statistics over the whole timerange
- Mark as New
- Bookmark Message
- Subscribe to Message
- Mute Message
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
I solved it myself!
sourcetype="Filter" a44 | transaction research maxevents=2 |stats min(duration) max(duration) avg(duration) median(duration) perc95(duration)
- Mark as New
- Bookmark Message
- Subscribe to Message
- Mute Message
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Sound like a good use case for the "transaction" command. Here's some documentation to get you started:
http://docs.splunk.com/Documentation/Splunk/5.0.3/Knowledge/Searchfortransactions
http://docs.splunk.com/Documentation/Splunk/5.0.3/SearchReference/Transaction
Make sure to take careful note of maxspan and maxpause options.
Hope this helps!
- Mark as New
- Bookmark Message
- Subscribe to Message
- Mute Message
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
A better alternative would be to find consecutive occurences of the same value for the keys "research" and "Locsite" (request and response)
