Infrastructure

What it does

When your agent uses it

• Onboarding a new cluster
• VM fleet to Grafana Cloud
• Picking the right Helm values for K8s scraping
• Wiring kube-state-metrics + node-exporter + cAdvisor

Example prompts

• “monitor my cluster”
• “send K8s metrics to Grafana”
• “scrape EC2 metrics”
• “/infrastructure”

Requirements

• Docker

Workflow steps

3 steps, taken from the step headings in SKILL.md.

1. 1Onboard a Kubernetes cluster (k8s-monitoring chart)
2. 2Monitor a Linux host
3. 3Pull AWS / Azure / GCP metrics

What it can do on your machine

Read from SKILL.md and the folder at commit 1ccacf2. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  Shell commands in SKILL.md call:

  • kubectl
  • helm
  • curl
  • jq

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  Hosts in commands or code, which the agent is likely to contact:

  • grafana.github.io

  Also links to:

  • grafana.com
  • github.com

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Infrastructure loads about 1.2k tokens when it runs, and up to ~2.5k if it reads all its reference files. Until then it costs about 207 tokens; SKILL.md has 157 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~207

When it runs · the whole SKILL.md, loaded when a task matches

~1.2k

With references · SKILL.md plus every file in references/, read only if the agent opens them

~2.5k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from grafana/skills at commit 1ccacf2, republished under its Apache-2.0 licence (© grafana). 157 words, ~1,188 tokens.

name

infrastructure

description

Ship Kubernetes, host, container, and cloud-provider telemetry into Grafana Cloud — `k8s-monitoring` Helm chart for K8s clusters (metrics + logs + traces + events + cost), Alloy `prometheus.exporter.unix` for Linux hosts, cAdvisor + Docker discovery for containers, and CloudWatch / Azure Monitor / Google Cloud Monitoring datasource setup. Use when onboarding a new cluster or VM fleet to Grafana Cloud, picking the right Helm values for K8s scraping, wiring kube-state-metrics + node-exporter + cAdvisor, alerting on `PodCrashLooping` / node memory / PVC capacity, or pulling AWS / Azure / GCP cloud metrics — even when the user says "monitor my cluster", "send K8s metrics to Grafana", "scrape EC2 metrics", "cluster pod logs", or "install the monitoring helm chart" without naming `k8s-monitoring` or Alloy.

license

Apache-2.0

# 1. Create the namespace + secret
kubectl create namespace monitoring
kubectl create secret generic grafana-cloud-secret \
  -n monitoring --from-literal=api-key=<your-api-key>

# 2. Install — values.yaml in references/k8s-monitoring-values.md
helm repo add grafana https://grafana.github.io/helm-charts && helm repo update
helm install k8s-monitoring grafana/k8s-monitoring \
  --version 4.1.4 -n monitoring -f values.yaml

# 3. Verify every pod is Running
kubectl get pods -n monitoring
# Expect alloy-*, kube-state-metrics-*, node-exporter-*, etc. all Ready.

# 4. Verify no error logs in the metrics/logs/traces Alloys
kubectl -n monitoring logs deploy/k8s-monitoring-alloy-metrics --tail=50 | grep -iE 'error|level=err' || echo "clean"

# 5. Verify telemetry landed in Grafana Cloud
#    PromQL on the metrics datasource (should be > 0):
#      sum(up{cluster="production-us-east"})
#    LogQL on Loki:
#      sum(count_over_time({cluster="production-us-east"}[5m]))

# 1. /etc/alloy/config.alloy — see references/clouds-and-hosts.md for the full block
prometheus.exporter.unix "host"  { rootfs_path = "/" }
prometheus.scrape         "node" { targets = prometheus.exporter.unix.host.targets
                                   forward_to = [prometheus.remote_write.cloud.receiver] }

# 2. Reload Alloy and verify the unix exporter is up
systemctl reload alloy
curl -s http://localhost:12345/api/v0/web/components | jq '.[] | select(.id|contains("prometheus.exporter.unix"))'

# 3. Verify in Grafana Cloud — open the "Node Exporter Full" dashboard (ID 1860)
#    and pick your host from the `instance` dropdown.

# 1. After provisioning, restart Grafana to pick up the file
# 2. Verify the datasource — Grafana → Connections → Data sources → "Test"
#    Expect "Successfully queried the CloudWatch metrics API" (or equivalent).
# 3. Confirm a query — Explore → datasource → metric e.g.
#    CloudWatch namespace AWS/EC2 metric CPUUtilization, last 1h.

© grafana, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 2 other files (references) in skills/grafana-cloud/infrastructure of grafana/skills.

• SKILL.md
• references/clouds-and-hosts.md
• references/k8s-monitoring-values.md

Open the folder on GitHubat commit 1ccacf2