The Euxdat project uses a platform for collecting, storing, and analyzing of the systems’ monitored metrics.
For such purpose, it is used the open-source Prometheus toolkit.
It allows adding optional components, such as intermediary gateways.
In particular, the Pushgateway component is already setup in Euxdat, because it allows batch jobs to expose their metrics to Prometheus.
The next figure shows the metric collection architecture. The components of the Exudat platform can send metrics through the HTTP protocol to a remote Prometheus system. We consider having a single source of metrics because it allows having a common format for and it simplifies the analysis tasks, as well as the implementation of web interfaces with Grafana.
The next figure below shows metrics from the preliminary tests of the data-mover component. In particular, those results show that the bandwidth between the computation centers used in Euxdat, and between them and an office connection. Notice that the datacenters have between 10 and 20 times better bandwidth than the commonly available in many offices and homes, even being when the datacenters located in different countries, In the case, of the experiments evaluated, the office is located in Stuttgart (Germany) as well as the HLRS center, and the other data center is located in France.