Anomaly Detection with Median Absolute Deviation

The Median Absolute Deviation (MAD) algorithm is a powerful anomaly detection technique that can be used to identify time series data points that are behaving differently from others. MAD is a widely used algorithm in DevOps Monitoring, where it enables Site Reliability Engineers (SREs) to quickly identify unhealthy containers, VMs, or servers and diagnose infrastructure problems. The algorithm works by calculating the median absolute deviation of each time series at a given timestamp, and then flagging points that have a large deviation from the median as anomalous. MAD is highly effective and efficient, but can be sensitive to anomalies, especially when dealing with large datasets. To mitigate false positives, SREs can use techniques such as adjusting the threshold value or grouping data by region. The MAD algorithm has been implemented in Flux, a powerful open-source query language for InfluxDB, which allows users to write custom anomaly detection algorithms and integrate them into their monitoring pipelines. By using MAD, organizations can improve the efficiency of their root cause analysis efforts, reduce mean time to resolution (MTTR), and honor Service Level Objectives (SLOs).

No tracked trend matches for this post yet.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.