Dealing with multiple sources of telemetry data is becoming more and more important, especially given the popularity of the microservices architecture. However, monitoring in such a situation becomes challenging as it is not humanly possible to keep track of thousands of counters, particularly when issues span across multiple nodes (e.g., microservices). Techniques are presented herein that address the problem of monitoring and analyzing high dimensional telemetry (comprising, for example, thousands of counters). Aspects of the presented techniques support a system that applies a dimensionality reduction algorithm, that is based on standard deviation and z-score statistical values, transforming data with a large number of dimensions into a two-dimensional array (such as, for example, a scatterplot). Application of the presented techniques enhances issue detection by a human observer and the automated generation of alerts.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.