LinkMesh

Search docs, blog and changelog

ENDE

OpAMP server

A self-hosted OpAMP server for your collector fleet

LinkMesh is the OpAMP server side, running inside your own infrastructure. otelcol-contrib connects with the standard opampsupervisor over one outbound port-443 connection; you push RemoteConfig, read EffectiveConfig and health, and manage rollouts and drift across the fleet. Self-hosted, standard upstream collectors, priced per collector.

Try LinkMesh as your OpAMP server →

25 Collectors free after a no-card registration · 5 without one How the free tier works

The OpAMP loop, self-hosted

The opampsupervisor on the host connects to your LinkMesh server over OpAMP and manages the otelcol-contrib subprocess. Configuration and health travel over that connection; telemetry flows separately, straight from the collector to your backends.

CONTROL PLANE · OpAMPLinkMesh Serveryour hosted instanceOpAMP Serverendpoint /v1/opamptransport WSS · port 443OTel Hostopampsupervisor + otelcol-contrib:4317 OTLP / gRPC:4318 OTLP / HTTP:8888 own metricsRemoteConfig · Connection SettingsEffectiveConfig · Health · StatusDATA PLANE · OTLPlogs · metrics · tracesYour Observability BackendsLoki · Tempo · Prometheus · Grafana Cloud · …

01

A self-hosted OpAMP server

Problem

OpAMP is a protocol, not a product. To use it you need a server on the other end of the connection — one you can run yourself, not a hosted endpoint your agents phone home to.

LinkMesh

LinkMesh is an OpAMP server you self-host. otelcol-contrib connects with the standard opampsupervisor over a single outbound port-443 connection; the server manages the agent from there. No proprietary agent, no vendor cloud in the path.

The OpAMP server side, running inside your own infrastructure.

How-to: enroll otelcol-contrib over OpAMP →

02

Remote configuration

Problem

Hand-editing a config file per host and restarting is how OpAMP’s whole reason for existing gets ignored. You want to push config from the server and have agents converge.

LinkMesh

LinkMesh renders configuration centrally and delivers it over OpAMP’s RemoteConfig message. Agents apply it and report back their EffectiveConfig and health, so the server always reflects what is actually running on each host.

Push once from the server; every agent converges and confirms.

How remote configuration works →

03

Effective-config & health reporting

Problem

RemoteConfig alone is half the loop. Without the agent reporting what it actually applied, you are pushing config into the dark.

LinkMesh

Every agent reports its EffectiveConfig hash and health status back over the connection. LinkMesh surfaces convergence, flags an agent whose reported hash diverges from what it was assigned, and shows the timeline of applies and drift.

A closed loop: you see the config land, and you see it drift.

How effective-config reconciliation works →

04

Fleet-scale management

Problem

One agent over OpAMP is a demo. Hundreds of agents, grouped by environment and rolled out to safely, is the actual job.

LinkMesh

Organize agents into groups, stage a config rollout to one group before promoting it, keep a live inventory of versions and status, and version every change in Git with one-click rollback — all over the same OpAMP connection.

OpAMP as an operational control plane, not a single-agent protocol demo.

How a collector is modelled in LinkMesh →

What the OpAMP support covers — and where the edges are

Supported

  • ✓otelcol-contrib via the upstream opampsupervisor (recommended path)
  • ✓RemoteConfig delivery from a self-hosted server
  • ✓EffectiveConfig + health / status reporting from the agent
  • ✓Connection settings and outbound port-443 transport
  • ✓Grouping, staged rollout, drift detection and Git-versioned rollback on top of the protocol

Worth knowing

  • •Grafana Alloy is managed over Alloy’s native remotecfg endpoint, not OpAMP — LinkMesh manages both, but Alloy does not connect as an OpAMP agent.
  • •OpAMP’s optional package/binary auto-update flow is not how LinkMesh ships collector upgrades; upgrades follow your own rollout, not agent self-update.
  • •LinkMesh is the server side. You still run standard upstream collectors — LinkMesh does not fork or replace them.

Stand up the server, enroll an agent

Install the LinkMesh server on a Linux VM, then point otelcol-contrib at it with the standard opampsupervisor and a scoped enrollment token. The agent connects outbound over port 443 and appears in the fleet.

1 · Install the OpAMP server on any Linux VM — Ubuntu / Debian

curl -fsSL https://artifacts.saas.opensight.ch/binaries/linkmesh-server/latest/linkmesh-server_latest_amd64.deb -o linkmesh-server.deb && sudo apt install -y ./linkmesh-server.deb

2 · Enroll otelcol-contrib over OpAMP with the opampsupervisor: onboarding guide →

RHEL / Rocky / AlmaLinux and the full walkthrough: Install guide →

If nothing changes

Left alone, none of this gets cheaper: the volume grows on its own, the data that already left cannot be recalled, and each agent added is one more to remove later.

Once it is running

  • A per-collector bill that a volume spike does not move.
  • Sensitive fields masked on the host, before anything leaves the network.
  • Every config change a diff you can review and roll back.
  • Backends you can swap, because nothing proprietary sits in the path.

Common questions

Is LinkMesh a self-hosted OpAMP server?

Yes. LinkMesh is an OpAMP server you run inside your own infrastructure. otelcol-contrib connects with the standard opampsupervisor over a single outbound port-443 connection, and the server manages the agent over the protocol.

Which collectors connect over OpAMP?

otelcol-contrib, via the upstream opampsupervisor. Grafana Alloy is managed too, but over its native remotecfg endpoint rather than OpAMP — LinkMesh manages both from one control plane.

Does LinkMesh support RemoteConfig and health reporting?

Yes. LinkMesh delivers configuration over OpAMP’s RemoteConfig message and receives EffectiveConfig and health back from each agent, so the server reflects what is actually running and flags drift when a reported config diverges from what was assigned.

Does OpAMP mean my telemetry flows through LinkMesh?

No. OpAMP carries management messages — configuration, health, and status — only. Your collector sends telemetry directly to your own backends; it never transits LinkMesh.

Does LinkMesh use OpAMP’s package auto-update to upgrade collectors?

No. Collector upgrades follow your own staged rollout rather than OpAMP’s optional agent self-update flow, so version changes are deliberate and observable across the fleet.

How is LinkMesh priced?

By managed collector, never by data volume. The first 25 collectors are free after a no-card registration in the OpenSight Customer Portal (5 without one); beyond that it is a flat USD 12.50 · CHF 12.00 · EUR 12.50 per collector per month, billed annually, and the bill stops growing at 200 collectors.

Run OpAMP on your own terms

A self-hosted OpAMP server with the fleet management built on top — grouping, rollouts, drift, and rollback. Standard upstream collectors, priced per collector. The first 25 Collectors are free after a no-card registration in the OpenSight Customer Portal (5 without one).

Start with one test collector.

Save the current configuration, connect one collector and verify a real signal in your backend. Test one change and its rollback before expanding to the fleet.

Setup guides and evaluation checklist

No email required for the download. Includes failure tests, rollback and checks for ingest-cost assumptions.