arrow_back Building Enterprise grade ML Apps : Tools and Architectures
Optimisation using Julia arrow_forward
Alerting @ AppDynamics: Simplifying User Experience for Data Intensive Applications
Submitted by Saarthak Puri (@saarp) on Friday, 12 April 2019
AppDynamics builds products that help large enterprises monitor their Application environments. A big part of monitoring is to be alerted when something goes wrong. AppDynamics provides tools that help users build these alerts, and over the last ten years, they have been using these tools to build alerts for mission critical applications.
This talk goes over the challenges of building a product that requires users to comprehend patterns in their data and then use those patterns to build meaningful alerts.
I first identify the different patterns that metrics can take and what kind of alerting is meaningful and relevant to each pattern. The talk then explores the complexity of modern day application architectures and shows why it is difficult for users to find the right metrics or combinations of metrics to be alerted on.
I then take a deep dive into the process that users undergo to define, fine tune and maintain alerts. I first illustrate the need for alert sensitivity feedback and talk through various explorations of how to achieve this. I also talk about the differences in a user’s mental model of metric correlation and actual data, and how we can bridge that gap. I then define alert noise and use that definition to segment alerts as well as to justify the need for user feedback on alert effectiveness. I also talk about grouping alerts into situations or incidents.
I end this talk with an exploration of various use cases that justify the use of Machine Learning (specifically anomaly detection) to augment manually configured alerts.
- AppDynamics overview
- Use cases for alerting
- Alerting Fundamentals
- Metric Patterns
- Types of alerts
- Alert Configuration
- User Flows
- Sensitivity Feedback
- Metric Correlation: Mental Model vs Reality
- Alert Noise
- Incident Management and Alert Grouping
- Anomaly Detection
Saarthak Puri is a Senior Product Manager at AppDynamics, the market leader in Application Performance Monitoring. He is responsible for the core Alerting Product there. Prior to AppDynamics, he was a Senior Product Manager at Capillary, where he led their Data Science Products. He has been an Entreprenuer in the past. Saarthak is passionate about Enterprise Software, and is deeply interested in building simple and elegant user experiences for complex enterprise workflows. He has a Bachelor of Technology in electrical engineering from IIT-Roorkee.