Over the last few years we have slowly and methodically been building out the ML based capabilities of the Netdata agent, dogfooding and iterating as we go. To date, these features have mostly been somewhat reactive and tools to aid once you are already troubleshooting.
Now we feel we are ready to take a first gentle step into some more proactive use cases, starting with a simple node level anomaly rate alert.
We have recently added a more detailed anomaly rate chart to Netdata that breaks out the overall node anomaly rate by type, this lets you more easily see what parts of your infrastructure might be experiencing an uptick in anomalies when you see the overall node anomaly rate increase.
How does Netdata's machine learning (ML) based anomaly detection actually work? Read on to find out!
The menu (on the overview or single node tab) now has an anomaly rate button built into it that, for the entire visible window or a highlighted time range, shows the maximum chart anomaly rate within each section.
Read on to learn more about this new feature!
We have recently extended the native machine learning (ML) based anomaly detection capabilities of Netdata to support all metrics, regardless on their collection frequency (update every).
Previously only metrics collected every second were supported, but now Netdata can run anomaly detection out of the box with zero config on metrics with any collection frequency.
This post will illustrate an example of what this means using Prometheus metrics (via the Netdata Prometheus collector) since they typically have a default collection frequency of 10 seconds.
We have been busy at work under the hood of the Netdata agent to introduce new capabilities that let you extend the "training window" used by Netdata's native anomaly detection capabilities.
This blog post will discuss one of these improvements to help you reduce "false positives" by essentially extending the training window by using the new (beautifully named) number of models per dimension configuration parameter.
Following on from the recent launch of our Anomaly Advisor feature, and in keeping with our approach to machine learning, here is a detailed Python notebook outlining exactly how the machine learning powering the Anomaly Advisor actually works under the hood.
Today we are excited to launch one of our flagship ML assisted troubleshooting features in Netdata – the Anomaly Advisor.
The Anomaly Advisor builds on earlier work to introduce unsupervised anomaly detection capabilities into the Netdata Agent from v1.32.0 onwards.