Observability with Prometheus and Grafana

Grafana Cloud (Managed Prometheus Metrics)

Grafana Cloud is Grafana Labs’ managed observability platform. Its managed Prometheus service, Grafana Cloud Metrics, is built on Grafana Mimir and accepts Prometheus remote_write or scraping via Grafana Agent. It provides a unified stack for metrics, logs (Loki), and traces (Tempo) inside the Grafana UI. Mimir supports high cardinality and large-scale Prometheus workloads while remaining fully PromQL-compatible.

• Free Tier & Pricing: Grafana Cloud offers a permanent free tier with 10,000 active series and 14-day retention, plus 50 GB logs and 14-day trace retention. The Pro plan ($19/month base) includes 10,000 series and 13-month retention, with extra series billed at ~$6.50 per 1,000 per month. Enterprise plans start around $25k/year with lower pricing per 1,000 series, dedicated stacks, and longer retention. Pricing is based on active series and retention rather than sample count. Excess scrape frequency above 6 samples/minute counts as additional series. Querying has no separate charge.

• Storage Retention: Free tier retention is 14 days; Pro offers 13 months; Enterprise can extend further. Grafana Cloud retains full-resolution raw samples for the entire retention period (no automatic downsampling). Enterprise deployments may use tiered storage or archiving when longer retention is needed.

• Query Language & Tools: Full PromQL support is provided via Grafana’s hosted UI. Complex PromQL queries benefit from Mimir’s optimized query engine. Alerting is handled through Grafana’s unified alerting system, which evaluates PromQL-based rules and routes notifications to various contact points. Prometheus API access tokens allow external tools or external Grafana instances to query Grafana Cloud Metrics.

• Visualization & Integration: Grafana Cloud includes a hosted Grafana instance with full plugin and dashboard support. Users can import existing Prometheus dashboards directly. Grafana Agent can collect metrics, logs, and traces and forward them to Grafana Cloud. Integrations provide pre-built dashboards and scraping configs for Kubernetes, Linux, AWS, and other systems. Metrics, logs, and traces can be correlated through exemplars and linked panels.