Detecting Anomalies in Logs with Time Series Analysis
summary
Claims that log-derived time series let teams catch cost spikes early and shows a working pipeline: Cloud Logging via log router -> BigQuery, aggregate per-minute record counts, train a time-series model (demo uses ARMA / ARIMA via BigQuery ML), extract anomalies, export CSV and plot in Google Sheets. Demo highlights a correlated ~700% billing jump and surfaces anomalous intervals; the pipeline runs end-to-end and flags predictive deviations rather than generic “examples.”
AI-written summary · about
chapters (11)
- 0:00Introduction and Welcome
- 0:27Speaker Introduction
- 0:51Common Cloud Cost Issues
- 3:06Proactive Cost Management Overview
- 4:17Benefits of Proactive Cost Management
- 5:32Challenges in Analyzing Cloud Logs
- 6:18Anomaly Detection Techniques
- 7:26Machine Learning Methods for Anomaly Detection
- 9:06Google Cloud's Time Series Solutions
- 11:02Real-World Use Cases
- 13:06Demo: Detecting Anomalies with BigQuery
related talks
43mBeyond Logs: Real-Time Anomaly Detection with Machine Learning
34mAnomaly Detection in Time Series from Scratch Using...
45mAnomaly Detection in Time Series
30mBetter data quality with ML & visualization
19mThe Rise of AI in Anomaly Detection
27mBeyond Logs: Real Time Anomaly Detection with Machine Learning
45mUsing time-series data to better understand our products & infrastructure
31mAnomaly Detection on OpenStack Logs using M.L.