Metrics are usually not persisted in the application but in another system that is part of the environment, such as external monitoring solutions. This simplifies the implementation of metrics; the enterprise application keeps the information in memory and emits the specified metrics. External monitoring solutions scrape and process these metrics.

There are several techniques that can be used to emit and collect metrics. For example, measures can be formatted into custom JSON strings and exposed via HTTP endpoints.

A monitoring solution that is part of the Cloud Native Computing Foundation, and, as of today, has huge momentum, is Prometheus. Prometheus is a monitoring and alerting technology that scrapes, efficiently stores, and queries time series data. It gathers data that is emitted by some service over HTTP in a specific format. Prometheus is powerful in scraping and storing data.

For graphs and dashboards for business-related information, other solutions can be built on top of this. A technology that works well with Prometheus and provides many possibilities for appealing graphs is Grafana. Grafana doesn't store time series itself but uses sources such as Prometheus to query and display time series.

The following screenshot shows an example of a Grafana dashboard:

The idea of dashboards provides visibility for business experts and combines relevant information. Depending on the requirements and motivations, coherent information is combined into graphs that provide overviews and insights. Dashboards provide the ability to query and customize time series representations based on the target group.