Overview
Deploy configurations to a fleet of collectors and manage them remotely.
Fleet management connects a configuration you built in Telflo to the collectors running in your infrastructure. Instead of copying YAML to each host, your collectors connect to Telflo, pull their configuration, and report their status.
You roll a configuration out to the fleet, watch each collector's health and status, detect drift from the version you approved, and roll back, all without logging into individual hosts.
Fleet management delivers configuration to your collectors and reads their status back. It is never in your telemetry data path: telemetry flows directly from your collectors to your backends.
The Fleet management section appears when fleet management is enabled for your account.
Fleets come in three types, chosen when you create the fleet. In a server fleet, each collector connects to Telflo itself, from a host or a container. In a Kubernetes bridge fleet, one bridge installed in the cluster connects to Telflo and manages the cluster's operator-managed collectors. A Kubernetes supervisor fleet runs supervisor-managed collectors in a cluster without the Operator; it is available in private preview. See Kubernetes fleets.
How it works
- Publish a configuration version. See Versions & publishing.
- Create a fleet. The fleet issues an install token and starts with one deploy group, named Default. See Create a fleet.
- Assign the published version to a deploy group on the fleet's Deploy Groups tab. Each group can run its own configuration version. A new fleet opens with a setup checklist that walks through these steps.
- Install collectors from the group's Install button: run the prebuilt collector image, or install a supervisor bundle on a host. For a configuration on OpenTelemetry Collector Contrib, the install dialog lists the
${env:NAME}variables the configuration needs, so you can include values in the command or bundle. For a collector that runs in Docker, declare the host paths the configuration reads as volume mounts on the deploy group first. See Install collectors. - Each collector connects to Telflo, pulls its assigned configuration, applies it, and reports the result.
- Track the fleet from its detail page. See Monitor a fleet.
The collectors run the collector distribution the assigned configuration was built for. Server fleet bundle installs and Kubernetes bridge fleets run configurations on OpenTelemetry Collector Contrib. A ClickStack configuration, available in private preview, deploys through a Kubernetes supervisor fleet or a server fleet's container image option; it cannot be installed from a supervisor bundle. See Key concepts.
For a Kubernetes fleet, steps 2 to 4 differ. You create either Kubernetes fleet type with a name only. In a Kubernetes bridge fleet, you then deploy one bridge into the cluster, and configurations are assigned per collector. In a Kubernetes supervisor fleet, available in private preview, you assign a published configuration to each role on the fleet's Deploy Groups tab and deploy each role with a Helm command. See Kubernetes fleets.
Creating the fleet, assigning the version, and getting the install command can also be done from the Telflo CLI or the API: telflo fleet create creates a fleet, telflo deploy assigns a published version to a deploy group, and telflo fleet install prints the install command. Monitoring stays in the app, and Kubernetes bridge fleets are managed in the app only.
Key terms
| Term | Meaning |
|---|---|
| Fleet | A named group of collectors managed together, as a server fleet or a Kubernetes fleet. |
| Instance | A single collector that has enrolled in a fleet. |
| Bridge | The component a Kubernetes bridge fleet installs in the cluster. It connects to Telflo and manages the cluster's collectors. |
| Deploy group | A named group of collectors in a server fleet or a Kubernetes supervisor fleet. Each group holds an assigned configuration version, an assigned variable group, and the volume mounts its container installs use, managed on the fleet's Deploy Groups tab. The API calls a deploy group a deployment. Every new server fleet starts with a group named Default; in a Kubernetes supervisor fleet, each group is a role: node, cluster, or a gateway pool. Kubernetes bridge fleets assign configurations per collector instead. |
| Install token | A credential that authenticates a fleet's collectors. It is delivered with the install command, or set on the host as the OPAMP_TOKEN environment variable for a downloaded bundle. |
| Served configuration | The configuration version Telflo has delivered to an instance. |
| Assigned configuration | The configuration version an instance is set to run: the version the instance is pinned to, if any, otherwise its deploy group's assigned version. |
| Drift | An instance whose served configuration version does not match its assigned version. |
Prerequisites
- A configuration with a published version. See Versions & publishing.
- Hosts where you can run a collector, on Linux (arm64 or amd64/x86_64) as a container or directly on the host, or a Kubernetes cluster for a Kubernetes fleet.
Next steps
Last updated on