ADAPT is a new regression detection feature that addresses the challenge of determining which metrics to compare and what static thresholds to use in an automatic way.

How does it work?

• ADAPT uses machine learning for anomaly detection across all metrics captured during performance tests, including both load metrics from load testing tools and resource metrics for the system under test.
• By using results from multiple previous test runs as a baseline, ADAPT can filter out variability, enhance benchmark accuracy, and reduce the number of false positives.
• The more test runs are analyzed, the better ADAPT becomes at flagging significant changes in recorded metrics, regardless of whether a requirement is defined.

Here’s why ADAPT is a game-changer:

Solves The Challenge

Manual identification of performance regressions and resource consumption issues during software testing is a cumbersome process. Without an automated solution, engineers are burdened with sifting through countless metrics, “finding the differences”, setting explicit checks and deviation thresholds on the usual suspects, subsequently, risking oversight and delayed detection of critical issues.

Why does it matter?

This problem is pervasive due to the complexity of modern software systems and the ever-increasing demand for performance optimization. Without timely detection, performance regressions can escalate into costly production issues, leading to dissatisfied customers, increased support overhead, and compromised brand reputation.

The solution

With Perfana ADAPT, teams gain an early warning system that swiftly flags anomalies in performance metrics, enabling them to proactively address regressions and resource bottlenecks. By automating the detection process, engineers save valuable time and effort, allowing them to focus on strategic optimizations rather than manual data analysis. Ultimately, this results in consistently high-performing software, delighted customers, and reduced operational costs.

With ADAPT, engineers:

✅ Spot differences with little effort

✅ Save time on analysis

✅ Get the difference between the test runs

✅ Configure SLO’s and thresholds for all metrics

✅ By using all of the metrics, easily spot the root cause of a given user-facing regression.

✅ Make the shift from reactive firefighting to proactive control with Perfana ADAPT.