grafana-dashboards-and-alertslisted

# Grafana Dashboards and Alerts Build production-quality Grafana dashboards and alerting rules following proven conventions from large-scale multi-cluster observability deployments. This skill covers the entire lifecycle: JSON model structure, template variables for portability, RED metrics PromQL patterns, unified alerting with label-based routing, and dashboard-as-code provisioning. ## When to use - Building new Grafana dashboards for services, infrastructure, or business metrics - Writing or debugging PromQL queries for RED (Rate/Error/Duration) metrics - Authoring Grafana unified alerting rules with proper labeling and routing - Configuring dashboard template variables ($datasource, cascading label_values, custom variables) - Setting up contact points (Slack, PagerDuty, webhook) and notification policies - Provisioning dashboards and alerts as code (file provider, JSON in git) - Visualizing Tempo distributed traces (service graph nodeGraph panels, span-metrics RED) - Refactoring dashboards for portability across environments or datasource instances - Reviewing dashboards for anti-patterns (hardcoded UIDs, broad regexes, missing annotations) ## Core conventions ### Datasource stack and selection Modern observability deployments typically use: - **Prometheus** (primary metrics) — often multiple instances (prod/staging/dev) - **Tempo** (distributed traces) — service graph + span-metrics (RED derived from traces) - **Pyroscope** (continuous profiling) - **Cloud-managed