Splunk Search

How to plot a delta timechart of average response time

gokadroid
Motivator

I have data like:

timestamp, serviceName, responseTime(in ms)

I want to plot the per minute delta of avg. responseTime (difference between avg responseTime yesterday vs today) by serviceName. Average is taken every minute. I want to observe only half an hour window.

Sample data:

03/17/2017 00:00:01 service1 242 
03/17/2017 00:00:02 service2 300
03/17/2017 00:00:03 service3 350 
03/17/2017 00:00:04 service1 280 
03/17/2017 00:00:05 service2 290 
03/17/2017 00:00:06 service3 300 
:
:
03/18/2017 00:00:01 service1 1242 
03/18/2017 00:00:02 service2 1300
03/18/2017 00:00:03 service3 1350 
03/18/2017 00:00:04 service1 1280 
03/18/2017 00:00:05 service2 1290 
03/18/2017 00:00:06 service3 1300 

Now,

- The avg(ResponseTime) of service1 for 03/17/2017 00:00 is (242+280)/2 = 261ms
- The avg(ResponseTime) of service1 for 03/18/2017 00:00 is (1242+1280)/2 = 1261ms
- Hence the delta avg(RespTime) for service 1 at 00:00 between yesterday and today is 1261-261 = 1000ms. It might also be negative 1000 if it was 1261 yesterday and 261 today.

I want to plot this delta by service name on a timechart for a window of last 30 minutes from now only. Please assist.

NOTE

- Services are more than three
- One service might get called mote than other service within a minute. So service1 might get called multiple times within a minute while chances are service2 might not be called at all within that minute.
- There is no sequence in which services are called (sample data makes it look like service1, 2 and 3 are in sequence)
0 Karma
1 Solution

woodcock
Esteemed Legend

Here is a run anywhere example (you will swap your base search and host for service and 1h for 1m😞

index=_introspection sourcetype=splunk_resource_usage 
| timechart span=1h avg(data.reads_kb_ps) AS HourlyAvgResponseTime BY host
| untable _time host HourlyAvgResponseTime
| eval hourmin=strftime(_time, "%H:%M")
| reverse
| streamstats current=f last(HourlyAvgResponseTime) AS prevHourlyAvgResponseTime BY hourmin host
| reverse
| eval delta=HourlyAvgResponseTime-prevHourlyAvgResponseTime

View solution in original post

0 Karma

woodcock
Esteemed Legend

Here is a run anywhere example (you will swap your base search and host for service and 1h for 1m😞

index=_introspection sourcetype=splunk_resource_usage 
| timechart span=1h avg(data.reads_kb_ps) AS HourlyAvgResponseTime BY host
| untable _time host HourlyAvgResponseTime
| eval hourmin=strftime(_time, "%H:%M")
| reverse
| streamstats current=f last(HourlyAvgResponseTime) AS prevHourlyAvgResponseTime BY hourmin host
| reverse
| eval delta=HourlyAvgResponseTime-prevHourlyAvgResponseTime
0 Karma

gcusello
SplunkTrust
SplunkTrust

Hi gokadroid,
see timewrap command (http://docs.splunk.com/Documentation/Splunk/6.5.2/SearchReference/Timewrap)

your_search 
| timechart avg(responsetime) AS responsetime count span=min 
| timewrap 1d align=now 
| sort -_time 
| head 30 
| eval diff=responsetime_latest_day-responsetime_1day_before
| table _time responsetime_latest_day responsetime_1day_before diff
| rename responsetime_latest_day AS Today responsetime_1day_before AS Yesterday diff AS Difference

Using 2 days as time period
Bye.
Giuseppe

0 Karma

gokadroid
Motivator

Thanks for the quick response but can you please provide computing the delta part of it?

0 Karma
Get Updates on the Splunk Community!

New in Observability - Improvements to Custom Metrics SLOs, Log Observer Connect & ...

The latest enhancements to the Splunk observability portfolio deliver improved SLO management accuracy, better ...

Improve Data Pipelines Using Splunk Data Management

  Register Now   This Tech Talk will explore the pipeline management offerings Edge Processor and Ingest ...

3-2-1 Go! How Fast Can You Debug Microservices with Observability Cloud?

Register Join this Tech Talk to learn how unique features like Service Centric Views, Tag Spotlight, and ...