Menu
Grafana Cloud RSS

Send metrics to Grafana Cloud

Caution

Grafana Alloy is the new name for our distribution of the OTel collector. Grafana Agent has been deprecated and is in Long-Term Support (LTS) through October 31, 2025. Grafana Agent will reach an End-of-Life (EOL) on November 1, 2025. Read more about why we recommend migrating to Grafana Alloy.

Metrics are numeric representations of data measured over time; such as the number of user logins, or the temperature of a device. Metrics tell you how much of something exists, such as how much memory a computer system has available or how many centimeters long a desktop is. In the case of Grafana, metrics are most useful when they are recorded repeatedly over time. This permits you to compare things like how running a program affects the availability of system resources.

Not sure where to go?
Answer a few questions and Grot will show you a helpful next step.

A deeper introduction to Grafana Cloud metrics

Compared to other metrics aggregation systems, Grafana Cloud provides:

  • Scalable querying and storage — as a service

    Grafana Cloud can support hundreds of millions of active series - two orders of magnitude more than Prometheus - so it will grow with your applications and infrastructure. There are no artificial limits or barriers, just true horizontal scalability.

  • A global view

    With Grafana Cloud, you can run queries over data from multiple data centers in a single place, enabling a unified global view of your metrics. Access Control Grafana Cloud provides authenticated access to control your dashboards and metrics. You can specify who has access to what.

  • Reliability and ease of performance

    Grafana Cloud is fully supported and operated by the developers and maintainers behind Prometheus. We operate and scale the platform for you, keeping it up to date with improvements and new features.

  • Easy-to-instrument whitebox monitoring

    Grafana Cloud Metrics is powered by a flexible NoSQL data model designed for monitoring modern infrastructure & microservices, paired with a powerful query language, PromQL.

Ways to connect your data to Grafana Cloud

Once you’ve created a Grafana Cloud account, the next step is to connect data to your Grafana Cloud stack. You can connect to your data using one of three primary methods, depending on what service or database is producing the data and how you want to use Grafana Cloud to work with that data.

Integrations

Integrations bundle Grafana Agent, tailored Grafana dashboards, and alerting defaults for common observability targets like Linux hosts, Kubernetes clusters, and Nginx servers. With Grafana integrations, you can get a pre-configured Prometheus and Grafana-based observability stack up and running in minutes.

Consider using a Grafana integration when:

  • You want to store data in your Grafana Cloud instances and do not want to manage your own data store
  • You do not have a strong preference for specific query languages such as PromQL, LogQL, and TraceQL
  • You want to get started quickly
  • You want to take advantage of the pre-built dashboards and alert rules that come with each integration
  • There are no restrictions on installing Grafana Agent on systems you want to monitor

There are several integrations that do not require installing the Grafana Agent on the host you would like to monitor. These integrations can connect directly to your SaaS application, such as the CloudWatch metrics integration for pulling metrics from AWS CloudWatch. The Metrics Endpoint integration allows you to automate the scraping of any Prometheus-compatible publicly accessible URL. Both of these examples allow you to forward scraped metrics for storage in Grafana Cloud without requiring any additional infrastructure.

For the complete list of Grafana integrations, refer to Integrations reference.

The following topics help you get started with integrations.

TopicDescription
Get started with monitoring using an integrationWalk through a complete example of installing an integration.
About integrationsView an overview of how to use integrations.
Install and manage integrationLearn how to install an integration, and how to update or remove it.

Send data from an existing observability deployment

Another option is to send your data from an existing observability deployment. For example, if you are already collecting observability data with Prometheus but want a single place to store, manage, and act on it, you might choose to deploy Grafana Agent manually, use Prometheus remote_write, or use another supported method of forwarding metrics to Grafana Cloud.

Consider sending data from an existing observability deployment when:

  • You only want to send a subset of your metrics data because:
    • You are building a “Proof of Concept” in Grafana Cloud for demonstration purposes
    • You don’t want to incur the expense of sending all data
    • You only want to send metrics data associated with certain teams or projects
  • You prefer to continue managing an existing observability deployment or not ready to move completely to Grafana Cloud managed platform
  • You prefer the redundancy (intentionally keeping the same metrics on self-managed and Grafana Cloud)

You can send the following observability data to Grafana Cloud:

Visualize externally-hosted metric data

Grafana Cloud data source connections enable you to visualize metric data where it’s stored, which means that the data is not sent to and stored in Grafana Cloud.

Consider using metrics data source connections when:

  • Metric ingestion costs for forwarding to Grafana Cloud are prohibitive
  • Compliance does not allow storing data off-premises
  • Custom configurations are required (such as alpha Prometheus features)
  • Custom retention is needed longer than 13 months
  • Your team prefers to manage the metrics environment
  • The metrics are stored and managed elsewhere (Amazon Managed Prometheus, GCP Managed Prometheus)
  • You prefer to continue with your own non-Prometheus metric data sources because moving to Prometheus or Mimir and learning PromQL is undesired

The following list provides links to data sources that support metrics: