How is AIOps machine learning done?

[Saeka]

Hello,

We have found AIOps to be instrumental in the automatic remediation of alerts and probable root cause analysis.

So my question is, which data does AIOps analyze and machine learn?
For example,

- the huge amount of data collected so far by ServiceNow

- data collected in one instance

[Community Alums]

Hi @Saeka ,

 Follow this link :https://www.servicenow.com/community/itom-blog/aiops-platform-onboarding-new-apps/ba-p/2340718

[Saeka]

Thank you for your reply.

When processing data through AIOps, can the criteria be shared with ServiceNow customers?

[Daniel Needles1]

At a high level AI excels at correlation of data points while humans excel at attaching meaning to the situation.  Like most AIOps the approach is to let AI determine clusters and then simplify the human interface to attach meaning.  For example, imagine a bunch of routers go down at once.  The human can recognize a non monitored power source is what is in common.  That is the high level.  This video goes into a bit more detail and nuance on how ServiceNow approaches it - https://www.youtube.com/watch?v=FzGvHx-FpiQ  One additional bit to add that the video didn't delve into. It is really scary even at medium size operations how temporal correlation alone is incredibly effective at finding these event clusters.

In a previous life for Moogsoft I wrote a whitepaper on the theoretical "left" and "right" brain approach to AIOps (wasn't coined at the time) that can be found here - https://www.linkedin.com/pulse/situational-awareness-cloud-openstack-aws-dan-needles/  (In case you are interested in where things are heading.)    NOTE Transforms (such as ChatGPT) are really leveling up at what point things are tokenized before being correlated.  This will make a huge impact in 2-3 years if not sooner.