Cut your Grafana Cloud bill by attributing spend to teams and reducing telemetry volume. Covers FOCUS-compliant billing dashboards, cost-attribution labels in Alloy, Adaptive Metrics (cardinality reduction), Adaptive Logs (drop/sample), Adaptive Traces (tail sampling), usage alerts, and an optimization checklist. Use when investigating a high Grafana Cloud bill, attributing observability cost to a team or service, reducing active series / log bytes / trace spans, or setting up usage / quota aler
Grafana Cloud Cost Management is a guided skill for cutting an observability bill by attributing spend to teams and reducing the volume of telemetry sent to Grafana Cloud. It targets the common problem of runaway metric, log, and trace costs: unbounded cardinality, noisy health-check logs, and untamed trace volume that silently inflate the invoice. Rather than a black box, it walks through concrete, reversible workflows grounded in Grafana's own Cost Management, Adaptive Metrics, Adaptive Logs, and Adaptive Traces features.
The skill covers four main workflows: attributing cost to a team or service by adding `team` and `project` external labels in Alloy configs; cutting metric cardinality by pulling and applying Adaptive Metrics recommendations (with a preview step and a documented rollback via the UI or a DELETE on the rule API); dropping noisy logs in Alloy with drop and sampling stages; and setting usage alerts before hitting quota using ready-to-paste Prometheus rules such as `MetricsUsageHigh` and `LogsIngestionHigh`. Reference files supply the full Adaptive signal snippets, cost-finding PromQL, and a billing-unit table. Each workflow includes verification queries so the user can confirm the change took effect and measure the reduction.
It is aimed at platform, SRE, and observability engineers who run a Grafana Cloud stack and need to investigate a high bill, identify which team or service is burning the most telemetry, or proactively guard against quota overruns. Prerequisites include a stack with Adaptive signals enabled, Alloy or Grafana Agent ingesting data with appropriately scoped API keys, and admin access to apply recommendations.
A Grafana Cloud stack with Adaptive Metrics/Logs/Traces enabled (visible under Cost Management), Alloy or Grafana Agent ingesting telemetry with an API key scoped for metrics:write/logs:write (and optionally traces:write), and admin access to the stack to apply Adaptive recommendations.
It adds external labels such as team and project to the Alloy metrics and logs configs, reloads Alloy, then groups by those labels in Cost Management. A reference file provides the full Alloy snippet.
Yes. Adaptive Metrics rules can be tested in Preview mode before applying and rolled back by disabling the rule in the UI or issuing a DELETE on the rule API. Log filters and sampling stages live in Alloy config and can be removed.
The skill's optimization checklist cites typical figures such as 40-60% series reduction from Adaptive Metrics recommendations, plus savings from dropping health/readiness probe logs and tail-sampling traces to 5-10% while keeping errors and slow spans. Actual results depend on your workload.
Yes. It includes ready-to-paste Prometheus alert rules that fire when metrics usage exceeds 80% of quota or log ingestion crosses a daily byte threshold, along with cost-finding PromQL for the top-cardinality metrics.
Quick Setup:
.claude/skills/Repository
grafana/skills