Skip to content
anythingai.space
AIOpsintermediate13 min · 9 steps

AIOps Techniques

The ML toolbox behind AIOps — time series, regression, clustering, classification, NLP for logs and deep learning — and which question each one answers.

Ops datalogs · metrics · eventsTime-series analysistrends · seasonalityClusteringgroup similar alertsClassificationlabel incident typeRegressionpredict a valueNLP for logsread log textDeep learningcomplex patterns
datarequestresponse
1 / 9
Key ideaStep 1 of 9

Six techniques, six questions

AIOps is not one algorithm. It applies different ML methods to ops data depending on the question: is this normal over time? (time-series analysis), what value comes next? (regression), which alerts belong together? (clustering), what kind of incident is this? (classification), what does this log line say? (NLP), and is this combination of many signals odd? (deep learning).

InsightPick the technique from the question, not from the hype. Most wins come from the simple ones.
What changed
  • +Six technique families

Lesson 3 of 5 in AIOps