What is Prometheus & Grafana?
Prometheus is an open-source monitoring system that uses a dimensional data model and a flexible query language. It provides an efficient time series database and supports modern alerting for applications, systems, and services.
Why use it in the API Lifecycle?
- Dimensional data model: Prometheus models time series in a flexible dimensional data model. Time series are identified by a metric name and a set of key-value pairs.
- Queries with PromQL: The PromQL query language allows querying, correlating, and transforming time series data for visualizations and alerts.
- Alerting Management: Alerting rules are based on PromQL and leverage the dimensional data model. A separate Alertmanager component handles notifications and silencing.
- Simple Deployment and Operation: Prometheus servers operate independently and rely on local storage. The statically linked binaries developed in Go are designed for easy deployment across various environments.
- Instrumentation Libraries: Prometheus offers official and community-contributed metrics instrumentation libraries that cover most major languages.
- System Integrations: Prometheus includes official and community-contributed integrations (exporters) to extract metrics from existing systems.
Best Practices
- Leverage Prometheus’s native integration with cloud-native environments like Kubernetes for continuous service discovery and monitoring.
- Configure a dedicated Alertmanager component to handle alert notifications and implement silencing rules effectively, separate from the core Prometheus server.
- Deploy Prometheus servers independently and rely on local storage for streamlined operation and simplified deployment across diverse environments.
- Utilize official and community-contributed instrumentation libraries to ensure standardized metric collection directly within applications written in various programming languages.
