topic Re: How to calculate average for the data which will get changed based on time interval ? in Splunk Search

How to calculate average for the data which will get changed based on time interval ?

georgear7 — Tue, 11 Aug 2020 16:10:11 GMT

I have below kind of data.

App Name Status
App1 0
App2 0
App3 0
App4 0
App5 0
App6 1
App7 0
App8 0
App9 0
App10 0

0 - Success
1 - Failure

Assign,
0 as 100%
1 as 0%

Here status value will get updated every 5 mins. So my requirement is to calculate the average from the starting time of day till present time by app. In this table, success percentage will be 90%. Similarly query should work if i want to calculate the average time after n'th update.

Next thing is i want to keep current status in one variable. and my final output should be having 2 data, one is current status and other one is Average data.

Re: How to calculate average for the data which will get changed based on time interval ?

georgear7 — Tue, 11 Aug 2020 16:12:47 GMT

@richgalloway @rnowitzki

Re: How to calculate average for the data which will get changed based on time interval ?

richgalloway — Tue, 11 Aug 2020 20:24:42 GMT

Here's a run-anywhere search for the first set of requirements.

| makeresults | eval _raw="App Name Status App1 0 App2 0 App3 0 App4 0 App5 0 App6 1 App7 0 App8 0 App9 0 App10 0" | multikv forceheader=1 ``` Above just defines test data``` | stats sum(Status) as status,count | eval SuccessRate= 100-((status*100)/count) | table SuccessRate

I don't understand the remaining requirements. What is the average of? Over what time?

Re: How to calculate average for the data which will get changed based on time interval ?

georgear7 — Wed, 12 Aug 2020 02:23:28 GMT

@richgalloway Thanks for your reply.

As i mentioned, Status of each App will get updated every 5 mins, It's live data.

For eg, if you take App1, starting from day...Means from 12.00 AM..

At 12.00 AM, first status will be updated for all the apps. then at 00.05, 00.10, 00.15........It goes on

Finally what my requirement is, to calculate the average and current status of each app separately by assigning 100 to success (ie. 0) and 0 to failure(ie. 1). I need this day wise, so next day it should start from 12 AM again.

If i run my query at anytime in a day, it should give me current status and average status for each app like below.

The final output table should be,

App	Current Status	Average Status
App1	100	90%
App2	100	100%
App3	0	80%
App4	100	100%
App5	100	100%

Hope this clarifies you.

Re: How to calculate average for the data which will get changed based on time interval ?

richgalloway — Wed, 12 Aug 2020 12:01:48 GMT

Try this

| makeresults | eval _raw="App Name Status App1 0 App2 0 App3 0 App4 0 App5 0 App6 1 App7 0 App8 0 App9 0 App10 0" | multikv forceheader=1 ```Above is just for test data``` | stats sum(Status) as status,count, latest(Status) as Current by App | eval Average= 100-((status*100)/count), Current=100-(Current*100) | table App, Current, Average