**Grafana** is the open source observability and data visualisation platform that serves as the primary dashboard layer for nearly every component of the open source hyperscaler stack. It connects to [[Prometheus]], [[Loki]], [[Tempo]], [[Jaeger]], and dozens of other data sources to provide unified operational visibility. --- ### First Principle: Data only has value when it's visible to the people who can act on it. Grafana's genius is treating every time-series database, log system, and trace store as a data source behind a uniform dashboard API. Operators don't learn a different UI for metrics, logs, and traces — they learn Grafana once and use it everywhere. --- ### Key Considerations - **Data Sources**: Native support for [[Prometheus]]/[[Mimir]]/[[Thanos]] (metrics), [[Loki]] (logs), [[Tempo]]/[[Jaeger]] (traces), Elasticsearch, InfluxDB, PostgreSQL, and many more. - **Dashboards**: Dashboards are JSON — version-controllable, shareable, importable from the Grafana community. Most major open source projects ship ready-made Grafana dashboards. - **Alerting**: Grafana Alerting provides a unified alert rule system across all data sources — supplementing or replacing [[Prometheus]] Alertmanager. - **Grafana LGTM Stack**: Grafana Labs maintains a tightly integrated stack: [[Loki]] (logs) + Grafana (visualisation) + [[Tempo]] (traces) + [[Mimir]] (metrics). - **Explore Mode**: Grafana's Explore interface allows ad-hoc queries against any data source — invaluable for incident investigation without building a dashboard first. - **Correlation**: Clicking a spike on a [[Prometheus]] metric graph can jump to the relevant [[Loki]] log lines from that exact time window. --- ### How It Fits ``` [[Prometheus]] (metrics) + [[Loki]] (logs) + [[Tempo]] (traces) → Grafana (unified visualisation layer) → Operations teams, SREs, on-call engineers ``` [[Prometheus]] | [[Loki]] | [[Tempo]] | [[Jaeger]] | [[Mimir]] | [[Thanos]] | [[Open Source Hyperscaler MoC]]