AI for DevOps (AIOps)

Applying AI/ML to operations: anomaly detection, root-cause, forecasting, and automation.

Extract signals from logs and classify errors with ML.

import re
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.linear_model import LogisticRegression

logs = ["ERROR DB timeout", "WARN cache miss", "INFO started"]
labels = [1, 0, 0]
vec = TfidfVectorizer()
X = vec.fit_transform(logs)
clf = LogisticRegression().fit(X, labels)
pred = clf.predict(vec.transform(["ERROR connection refused"]))
print(pred)

# Pseudocode: scale when forecasted CPU > threshold
# fetch metrics → forecast → call cloud API to scale

Build a Python service that ingests logs, extracts features, trains a classifier, and exposes an API to flag anomalous messages.