Outlier detection

Hawkeye detects outlier through the following options:

• Machine learning by using Auto Learned Threshold: Thresholds are the baselines to decide if tests are passing or failing. Set up correct thresholds to understand the test results. Auto Learn Threshold helps to calculate threshold values. Auto Learn Threshold reads the metric data for a specific scheduled test case in set of batches and calculate threshold to be set for the metric.
• Configuring Outlier per test: Often, a few isolated failed cases are normal and there is no reason to issue an alert. Hawkeye detects an outlier if a set of failed cases occurred in a configured time frame.