Cloud provider observability fundamentals on Grafana Labs

Monitor cloud providers in Grafana Cloud

Fri, 02 Oct 2026 08:47:49 -0500

Monitor your cloud providers in Grafana Cloud

Your workloads run in AWS, Azure, or GCP, and each provider keeps its metrics and logs in its own monitoring service. Cloud Provider Observability is the Grafana Cloud app for monitoring those services from one place. It fits a single provider as well as all three.

This journey shows you two ways to work with that data in Grafana Cloud:

Send to and store it in Grafana Cloud with Cloud Provider Observability and use prebuilt dashboards and alerts.
Query it in place with a data source. The CloudWatch and Azure Monitor data sources ship importable dashboards; you build alerts yourself.

Either way, you can visualize, alert on, and correlate your cloud data alongside everything else you monitor.

What you’ll learn

How to collect metrics from AWS, Azure, and GCP
How to collect logs from your cloud services
How to decide between storing your cloud data in Grafana Cloud and querying it in place with a data source
How to decide which collection method fits your needs

Journey duration

~35-45 minutes

Two options

Fri, 02 Oct 2026 08:47:49 -0500

Two options for monitoring cloud provider data

Both options let you visualize, correlate, and alert on your cloud data. The difference is where your data lives and how much comes prebuilt.

Cloud Provider Observability

The Grafana Cloud product for monitoring AWS, Azure, and GCP. It sends your metrics and logs to Grafana Cloud Metrics and Grafana Cloud Logs, so:

Prebuilt dashboards and alerts: Cloud Provider Observability installs dashboards and alerts for AWS, Azure, and GCP.
One query language: query metrics and logs with PromQL and LogQL.
Added cost: storing data in Grafana Cloud adds ingestion and storage cost.

A data source

A CloudWatch, Azure Monitor, or Cloud Monitoring data source connects to your data where it lives, so:

No Grafana Cloud ingestion or storage cost: your data isn’t copied or stored again. Your provider may bill each API query.
Importable dashboards: the CloudWatch and Azure Monitor data sources ship dashboards you can import. They don’t ship alert rules. The Google Cloud Monitoring data source ships neither.
Occasional access: a good fit for data you query occasionally or would rather not duplicate.

The practical distinction is cost:

Cloud Provider Observability adds ingestion and storage cost for a prebuilt, curated experience.
A data source adds no Grafana Cloud ingestion or storage cost because it queries your data where it lives. Your provider may bill each API query.

You don’t have to choose one approach for everything. For example, store your production data in Grafana Cloud, and query dev or ad-hoc data in place with a data source.

Prerequisites

Fri, 02 Oct 2026 08:47:49 -0500

Prerequisites

To follow along with the path exercises, you need:

A Grafana Cloud account (sign up free)
Access to at least one cloud provider (AWS, Azure, or GCP), with running resources to monitor such as EC2 instances, Azure VMs, or GCE machines

The permissions you need depend on which path you take.

Requirements for Cloud Provider Observability

An Admin or per-provider Writer role in Grafana Cloud to configure Cloud Provider Observability
Cloud provider permissions to create IAM roles or service accounts for data collection
Without a Grafana Cloud basic role, the Datasources:Reader fixed role to see dashboard data
With no basic role or only the Viewer basic role, the Alerting:Rules Writer, Alerting:Set provisioning status, and Folders:Writer fixed roles to install alerts and access folders

Note
Cloud Provider Observability doesn’t support AWS PrivateLink or Azure Private Link.

Requirements for using a data source

Permission to add a data source in Grafana
Read-only credentials for your cloud provider (no Cloud Provider Observability write access or access policy tokens needed)

Recommended prior knowledge

Topic	Why it helps
Basic Grafana navigation	You’ll explore dashboards and data sources
Your cloud provider’s console	You’ll configure IAM and collection jobs

You can also complete the journey without hands-on practice. The concepts stand on their own.

Decide what you want to learn about

Fri, 02 Oct 2026 08:47:49 -0500

Decide what you want to learn about

Start with whichever area fits, or click Next to go through the journey in order.

Cloud metrics: Numeric measurements from your cloud resources, such as CPU, memory, and request counts. A good place to start, since metrics are quick to set up and drive dashboards and alerts.
Cloud logs: Event records such as errors, audit trails, and access logs, for troubleshooting and investigation.
Query with data sources: Query your cloud data where it lives instead of storing it, when you’d rather avoid the extra ingestion and storage cost or only need occasional access.