A collection of Grafana dashboards based on Coroot node agents and Prometheus-compatible data sources.
Focus on observability with simple deployment — no UI, no ClickHouse, only metrics.

1. Create VMRule based on the VMRule.yaml (or use as Prometheus recording rules).
2. Import the dashboards into your Grafana.
3. Set up your Prometheus (or VictoriaMetrics) as the data source.
4. Use and enjoy.

👉 Demo (YouTube)

1. Coroot Node Agents (no UI, no ClickHouse, only node agents).
2. Prometheus or any compatible system (VictoriaMetrics recommended).
3. Recording rules support (documentation).
4. Grafana.
5. Eyes, hands, and care for observability.
6. Optionally: kube-state-metrics and cAdvisor for extended container metrics.

• Drilldown: from an overview to namespace/service level.
• TCP metrics and full connection details including FQDN and traffic direction.
• HTTP metrics: errors, latencies, and traffic.
• Postgres and Kafka metrics (beta).
• CPU throttling metrics for better root cause analysis (RCA).
• IN/OUT separation: errors and latencies by traffic direction.

• Add error metrics for TCP connections.
• Redis and other stateful workloads.
• Instance/node-level dashboards.
• Visual graph support for stateful workloads.
• Error timeline improvements.
• PSI (Pressure Stall Information) metrics integration.

Minimalistic setup: only agents and metrics. No ClickHouse. No heavy backend.