Add a collector

A “collector” is the data plane — the OpenTelemetry runtime (Grafana Alloy or otelcol-contrib) that actually receives, processes, and ships your telemetry. LinkMesh is the control plane that configures it. This page helps you pick how to connect one.

The collector runtime is always upstream

Whichever runtime you choose below, the collector is upstream Grafana Alloy or upstream otelcol-contrib, installed from its own packages. LinkMesh ships no collector distribution of its own.

Pick a runtime

A collector runs one of two runtimes, and you deploy it on one of two substrates (a VM or a Kubernetes cluster). The runtime decides how the collector receives its config:

Runtime	managementMode	How config arrives	Best for
Grafana Alloy + remotecfg	alloy-remotecfg	Collector pulls config over Bearer-authenticated HTTPS (remotecfg)	You run / prefer Alloy as a single self-contained binary
otelcol-contrib + OpAMP	opamp	Server pushes config over OpAMP (WSS); opampsupervisor applies it	You’re standardising on the OpenTelemetry OpAMP ecosystem

Both give you central config push, fleet status, and per-component throughput on the topology canvas. They’re alternatives, not stages — pick one per collector (you can mix runtimes across a fleet).

Check the supported version window first

Whichever runtime you pick, install a version inside LinkMesh’s supported window — a collector below the floor rejects its whole config and runs nothing, while still reporting as online. The current windows for both runtimes are on Compatibility & supported versions. Check it before you install, not after.

Grafana Alloy + remotecfgInstall upstream Alloy, add a remotecfg block, paste a token.

otelcol-contrib + OpAMPInstall upstream otelcol-contrib, add an OpAMP extension + supervisor.

Optional: + onboarding agent

Either runtime can be paired with linkmesh-agent, an optional edge connector that discovers the services, log sources, and Kubernetes workloads on a host and feeds them into the onboarding wizard. It is not a third path or a tier — it never installs, supervises, or configures the collector, and carries no telemetry. Add it only when discovery saves you onboarding work. See The agent.

Verify (either runtime)

On the host:

# Alloy runtime
systemctl is-active alloy
# otelcol + OpAMP runtime (the supervisor manages the collector process)
systemctl is-active otelcol-contrib    # or your opampsupervisor unit

In the LinkMesh UI, open Collectors — the host appears with status Online/Connected and its runtime (alloy-remotecfg or opamp). Click in to see runtime, last-seen, and live throughput.

Next steps

• Compatibility & supported versions — the supported runtime windows, and which agent versions a server tolerates.
• Build your first pipeline — wire the new collector through a processing pipeline to a destination.
• Enrollment tokens — TTL, scope, regeneration, audit visibility.
• Troubleshooting enrollment — when a host doesn’t show up.