Sending telemetry data to Grafana Cloud — metrics via Prometheus remote write or OTLP, logs via Loki push or Alloy, traces via OTLP to Tempo, profiles via Pyroscope. Covers Alloy-based pipelines, direct SDK/agent integrations, cloud integrations catalog, and credentials management. Use when connecting an application or infrastructure to Grafana Cloud, setting up data ingestion, configuring remote write, or choosing between ingestion methods.
Send Data to Grafana Cloud is a reference skill for connecting applications and infrastructure to Grafana Cloud across all four telemetry signals: metrics, logs, traces, and profiles. It addresses the common friction of getting data flowing — knowing which endpoint, which credential, and which ingestion method to use — by laying out the credential fields for each signal and then providing copy-ready configuration for the main integration paths.
The skill's centerpiece is a full Alloy configuration covering all signals: Prometheus scrape and remote_write for metrics, Loki file source and write for logs, and an OTLP receiver forwarding traces to Tempo, all authenticated via environment-variable-backed basic auth. It also documents direct Prometheus remote_write in prometheus.yml, a direct Loki push via curl, OpenTelemetry SDK export configured through OTEL environment variables with a Python example, the cloud integrations catalog (AWS CloudWatch, Azure Monitor, Kubernetes, and many databases and runtimes) installed via UI or API, a Kubernetes Helm chart values file for the k8s-monitoring operator, and API-key creation with ingestion-specific roles such as MetricsPublisher and LogsPublisher. Credentials throughout are handled through environment variables and Kubernetes secrets rather than hard-coded values.
It is aimed at developers and platform engineers onboarding a service or cluster to Grafana Cloud, setting up remote write, or choosing between Alloy, direct SDK/agent, and pre-built integration approaches. The main prerequisite is a Grafana Cloud stack whose per-signal credentials are found in the portal under My Account, Stack, Details.
The skill recommends Alloy for all signals as the default. It also documents direct Prometheus remote_write, direct Loki push, OpenTelemetry SDK export, pre-built cloud integrations, and a Kubernetes Helm chart, so you can pick based on your stack.
In the Grafana Cloud portal under My Account, Stack, Details. Each signal (metrics, logs, traces, profiles) has its own URL, username, and password/API key.
Through environment variables (for example GRAFANA_CLOUD_API_KEY read via sys.env in Alloy and OTEL_EXPORTER_OTLP_HEADERS for the SDK) and Kubernetes secrets in the Helm values, rather than being written inline.
Yes. It includes a values.yaml for the grafana/k8s-monitoring Helm chart that enables metrics, pod logs, and traces, plus the helm install command.
The skill shows a curl call to the Grafana auth keys API and lists the data-ingestion roles: MetricsPublisher, LogsPublisher, TracesPublisher, and ProfilesPublisher.
Quick Setup:
.claude/skills/Repository
grafana/skills