Geometric-aware distance measure for performance testing analysis

A system comprising:

• at least one processing device comprising a processor coupled to a memory;
• the at least one processing device being configured:
• to transform, by a monitoring system, time series data into graph representations by creating nodes corresponding to points in the time series data and connecting the nodes with edges based on visibility between corresponding points in the time series data;
• to extract, by the monitoring system, graph features from the graph representations;
• to compute, by the monitoring system, a hybrid distance metric by combining Sum of Squared Distances (SSD) and Fréchet Distance;
• to generate, by the monitoring system, a similarity score based on the hybrid distance metric, wherein the monitoring system compares the similarity score against a baseline performance curve to detect deviations indicative of performance anomalies in a computing system;
• and
• automatically adjusting, by the monitoring system, computing system resources based on the detected performance anomalies.

A computer program product comprising a non-transitory processor-readable storage medium having stored therein program code of one or more software programs, wherein the program code when executed by at least one processing device causes said at least one processing device:

• to transform, by a monitoring system, time series data into graph representations by creating nodes corresponding to points in the time series data and connecting the nodes with edges based on visibility between corresponding points in the time series data;
• to extract, by the monitoring system, graph features from the graph representations;
• to compute, by the monitoring system, a hybrid distance metric by combining Sum of Squared Distances (SSD) and Fréchet Distance;
• to generate, by the monitoring system, a similarity score based on the hybrid distance metric, wherein the monitoring system compares the similarity score against a baseline performance curve to detect deviations indicative of performance anomalies in a computing system;
• and
• automatically adjusting, by the monitoring system, computing system resources based on the detected performance anomalies.