Slide 2 of 8

How Fleet Management works

How it works

Collectors register with Fleet Management, receive remote configurations, and send health telemetry back to Grafana Cloud.

Fleet Management architecture showing collectors registering with Fleet Management, receiving remote configurations, and sending health telemetry to Grafana Cloud

The flow

  1. Register — Collectors include remotecfg block in local config
  2. Discover — Collectors appear in Fleet Management inventory
  3. Configure — Create pipelines and assign with attribute matching
  4. Distribute — Collectors poll for matching configurations
  5. Monitor — Collector health telemetry flows to Fleet Management dashboards

Key concepts

ConceptWhat it means
CollectorAn agent registered with Fleet Management
Configuration pipelineA piece of config that can be assigned remotely
AttributeKey-value pair used to match pipelines to collectors
remotecfg blockLocal config that registers a collector with Fleet Management

Local + remote configs

The collector runs local and remote configurations in parallel, isolated from each other. This means:

  • Remote config errors don’t break local config
  • You can mix and match based on your needs
  • Gradual migration from local to remote is safe

Script

Here’s how Fleet Management works under the hood.

It starts with your collectors. Each collector has a local configuration file that includes a special block called remotecfg. This block registers the collector with the Fleet Management service, telling Fleet Management “hey, I exist, I’m allowed to talk to you, here are my attributes, and I’m ready to receive remote configuration.”

Once registered, the collector appears in your Fleet Management inventory in Grafana Cloud. You can see its health status, version, operating system, and any attributes you’ve assigned, like environment, region, or team.

Here’s where it gets powerful. In the Fleet Management interface, you create configuration pipelines. A pipeline is a piece of configuration, maybe a metrics scrape or a processing rule. You assign pipelines to collectors using attribute matching. For example, “apply this pipeline to all collectors where environment equals production.”

The collectors periodically poll the Fleet Management API (by default every 60 seconds) and pull down any matching pipelines. These remote configurations are run in parallel with their local configs. If a remote config has an error, the collector falls back to the last known good configuration, so you don’t accidentally break your telemetry.

The result? You define configurations once, assign them matching attributes, and Fleet Management handles distribution to potentially hundreds or thousands of collectors.

Previous Next