Adaptive Alerts is a new feature from Rollbar that adds to our reliable, informative and actionable alerts about unexpected issues in monitored applications and services. Adaptive Alerts uses anomaly detection to learn the standard behavior of enterprise applications, and alerts developers about atypical exception rates, reducing unwanted noise.
When Errors Become More Active
Enterprise applications often have a large number of unique errors that are open, but not a priority to fix as long as they occur at a low rate. Developers and SREs want to know when such errors become more active than usual, so that they can resolve them if needed. But alerts based on fixed error thresholds generate too much noise and waste valuable developer time.
Brian Rue, our CEO and co-founder says, “Error alerting is important, but historically developers have had to deal with a lot of noise. Adaptive Alerts is the next generation of trendline alerting in Rollbar. Compared to the previous generation, which Rollbar customers know as the 10^nth Occurrence and High Occurrence Rate, Adaptive Alerts sends 86% fewer notifications, thanks to automatically adjusting thresholds, and a broader exception-level view that effectively detects application-level trends.”
Training and Anomaly Detection
The Rollbar algorithm is continuously trained using two weeks of historical data for each product/application. After analyzing a project’s historical data, Adaptive Alerts determines a baseline, monitors the exceptions rate continuously, and sends an alert within minutes when it detects something significant. Users can adjust the minimum exception rate to alert and sensitivity to fine-tune what constitutes an anomaly.
Ethan Salisbury, principal software engineer at health and wellness company Plexus Worldwide, commented that, “Adaptive Alerts grouped errors together in a way that I couldn’t have.”
Adaptive Alerts is automatically available to Rollbar’s Advanced and Enterprise plan customers.