Industrial Self-Healing Measurement Systems

Abstract

Automated measurement programs (i.e., placeholders for large number of measurement systems) are an efficient way of collecting, processing, and visualizing measurements in large software development companies. The measurement programs rely both on the software for data collection, analysis, and visualization—measurement systems—and humans for reporting of the data, design, and maintenance of the measurement systems. As the outcome of the measurement program—visualized measurement data—is an important input for decision making in the companies, it needs to be trustworthy and up to date. In this paper we present an experience report on development, deployment, and use of a self-healing measurement systems infrastructure at Ericsson AB. The infrastructure has been in use for a number of years and handles over 4,000 measurement systems in a fully automated way. Monitoring and self-healing of the infrastructure lead to the availability of measurement systems 24/7 and reducing the costs of managing them.