Back in the dark ages of Elasticsearch, understanding why a query was slow required experience, intuition, reading some tea leaves and making sure the moon was

In the final article of this three-part series, we build a fully automated anomaly detector using Watcher to send email alerts.

In part 2 of this series on building a statistical anomaly detector, we graph the output in Timelion, a Kibana app for graphing time-series data.

In this series, we implement a statistical anomaly detector to detect when correlated changes in variance occur, providing forewarning of a disruption.

In Part 2, the Control Chart is used in tricky scenarios such as constantly increasing linear trends or cyclic/seasonal data.