All Apps and Add-ons

How to get prediction for next few days for certain categories?

pramit46
Contributor

I have disk utilization log for which i want to do prediction for next 3 days for the following levels:
location, client, server

Problem is timechart (which is required for the predict command) does not allow more than one field. How do i proceed?
I am using ml app.
My search command is something like this:

Search data| apply model|timechart avg(usage) as avg by location|predict avg future_timespan=3

How do i incorporate client and server in this so that i get prediction for each level?

0 Karma

DalJeanis
Legend

Basically, you can only do one thing at a time. However, you can fake out the system

 Search data
| apply model
| eval mygroupfield=mvappend("L=".location,
    "L=".location."!!!!C=".client,
    "L=".location."!!!!C=".client."!!!!S=".server)
| fields - location client server
| mvexpand mygroupfield
| timechart avg(usage) as avg by mygroupfield
| predict avg future_timespan=3

After that, you can filter or break up the predictions as you choose using something like this:

| rex mygroupfield  "L=(?<location>.+?)(!!!!C=(?<client>.+?))?(!!!!S=(?<server>.+?))?$"

There will be a single record for each location, a record for each location-client, and a record for each location-client-server.

0 Karma

pramit46
Contributor

thanks a lot. but the predict command says cannot find the field avg. when I look into the timechart, I see due to breakup, it shows all the values of mygroupfield but no specific group called avg or avg(usage).
any idea what I can do for this?

0 Karma
Get Updates on the Splunk Community!

Stay Connected: Your Guide to January Tech Talks, Office Hours, and Webinars!

What are Community Office Hours? Community Office Hours is an interactive 60-minute Zoom series where ...

[Puzzles] Solve, Learn, Repeat: Reprocessing XML into Fixed-Length Events

This challenge was first posted on Slack #puzzles channelFor a previous puzzle, I needed a set of fixed-length ...

Data Management Digest – December 2025

Welcome to the December edition of Data Management Digest! As we continue our journey of data innovation, the ...