The bill follows the volume
A single noisy deployment can move an ingest invoice, and nothing in the path lets you decide what was worth sending before you are charged for it.
Manage your OpenTelemetry and Grafana Alloy fleet centrally.
Telemetry bills grow with volume, not with value. A self-hosted control plane for your OpenTelemetry collectors — filter, mask and route telemetry before it is billed, priced per collector.
Self-hosted, free to start. 25 Collectors free after a no-card registration · 5 without one — How the free tier works.
Early Access
Work directly with the builders and lock in launch pricing.
Prefer to be guided in, or want product updates? Get early access:
You're on the list — we'll be in touch by email.
We only email you about early access and product updates — no newsletter, and your address goes to nobody else. Unsubscribe any time. Privacy.
No collector yet? Install LinkMesh with one command
Install 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.debRHEL / Rocky / AlmaLinux, collector fleets on Kubernetes, and the full walkthrough: Install guide →
Already running a collector? Connect it in ~10 lines
otelcol-contrib · OpAMP
Running otelcol-contrib?
Add the OpAMP extension and a token. ~10 lines of YAML. →
Grafana Alloy · remotecfg
Running Grafana Alloy?
Add a remotecfg block and a token. Upstream Alloy, no fork. →
Either way, your telemetry stays on your infrastructure. Optionally add the LinkMesh agent for guided onboarding — it auto-discovers services and log sources.
Volume decides your bill. You never chose the volume.
Every deployment adds telemetry. Priced per gigabyte, the cost of watching a system scales with the system itself — never with how much of that data anyone reads.
Sensitive fields leave first
Card numbers and addresses reach a third-party backend inside ordinary logs, and the record of what left is then someone else’s to produce.
Every agent is a one-way door
A proprietary agent per vendor means the next backend decision is a fleet-wide migration you pay for in engineering time.
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.
Explainer · 1:33
LinkMesh in ninety seconds
What a collector is, where the control plane runs, and why you pay per collector instead of per gigabyte.
LinkMesh — the ninety-second explainer
Already sold and want to see it installed? Watch the full install walkthrough →
Connect your first collector in three steps
Install the control plane
Choose a supported Linux host in your infrastructure and follow the installation guide.
Connect one collector
One config block on stock otelcol or Grafana Alloy, over OpAMP or remotecfg. There is no new agent to adopt.
Apply a change and verify it
Apply one configuration change, check collector status, and verify the resulting telemetry in your backend before expanding.
What is true 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.
Optional
Add the onboarding agent for a faster start
A plainly-enrolled collector works on its own. The optional LinkMesh agent is an edge connector that auto-discovers what's running on a host or cluster and turns it into ready-to-approve pipelines — so onboarding is minutes, not manual YAML. It carries no telemetry and never sits in the data path.
| Onboarding | Collector onlyOpAMP or remotecfg | + LinkMesh agentoptional edge connector |
|---|---|---|
| Service auto-detection | Author receiver config by hand | Detects running services + suggests receivers |
| Log-source discovery | Find log paths yourself | Guided picker + template library |
| Kubernetes workload rollup | — | Discovers pods / workloads for onboarding |
| Reachability preflight | — | Flags connectivity + unreadable sources up front |
| Host footprint | Your collector only | + a small agent (package or DaemonSet) |
First time? Walk through the full setup with screenshots on linkmesh.io/docs/quickstart.
One tool for the whole pipeline
Open standards and no lock-in, per-collector pricing, on-prem and PII-aware — one control plane for the four jobs every telemetry pipeline needs.
Your sources
Collect
Logs, metrics, and traces from any source, over standard OpenTelemetry Collectors — no proprietary agent, no fork.
Refine
Drop, sample, and aggregate noisy telemetry before ingestion. Measure the reduction against your own workload.
Protect
Mask credit cards, emails, and custom fields on the host, before anything leaves your network.
Route
Fan out to the right backend by label, environment, or tenant — Grafana, Splunk, Elastic, Azure, any OTel destination.
Any OTel destination
For the technical model behind these — Collectors, Sources, Pipelines, Routes, Destinations — see Concepts in the docs.
Who you are trusting with the pipeline
Nobody decides to pay by the gigabyte for their own logs. It arrives one deployment at a time, and by the time it reaches an invoice the data has already left.
Four things you can check before you install anything.
A named team, not a vendor queue
LinkMesh is built by OpenSight, an independent Swiss company founded by Roman Hüsler. A question about the architecture reaches an engineer.
We cannot see your telemetry
The control plane runs in your infrastructure. We have no access to your telemetry, your collector configurations or your dashboards — not as a policy, as an architecture.
Standard OpenTelemetry, so you can leave
Stock collectors over OpAMP and remotecfg. No proprietary agent and no fork, which is the same thing as saying there is nothing to unpick if you stop.
We publish what we do not have
No SOC 2, no ISO 27001, and a public list of the product’s current limits. You should not have to discover either of those during a procurement review.
Start where your problem is
Cutting an ingest bill, managing a collector or Grafana Alloy fleet, or keeping telemetry inside a boundary — go straight to the page for the job.
OpenTelemetry Fleet Management →Manage every collector from one place.Reduce Splunk Ingest Cost →Filter before the data reaches Splunk.Grafana Alloy Fleet Management →Manage Alloy centrally over remotecfg.Self-Hosted Control Plane →On-prem & air-gapped — telemetry stays put.PII Filtering & Redaction →Mask sensitive telemetry before egress.OpAMP Server →A self-hosted OpAMP server for your fleet.
Comparing tools? See LinkMesh vs Bindplane, Cribl, Grafana Fleet Management →
Replacing something? Alternatives to Bindplane, Cribl, Splunk forwarders and three more →
25 Collectors free for evaluation.
Test enrollment, pipeline configuration, PII masking, and multi-destination routing in your own environment — no volume-based cost risk during the trial. Beyond the Community limit, pricing scales by managed Collector.
25 Collectors free after a no-card registration · 5 without one — How the free tier works.
Get in touch
Questions about pricing, running a bigger fleet, or self-hosting? Tell us what you need and we'll get back to you.
Thanks — message sent.
We will reply by email to arrange the next step.
From the blog
Telemetry pipelines, OpenTelemetry, and observability cost control — from the team building LinkMesh.
LinkMeshOpenTelemetryObservability
Build a Central Observability Platform
linkmesh.io
Plan a central observability platform across on-premises and Azure: ownership, Collector blueprints, change control and measurable pilot acceptance criteria.
LinkMeshOpenTelemetryMigration
One Agent to Rule Them All
linkmesh.io
A forwarder, a Datadog Agent and a OneAgent on every host. How to consolidate monitoring agents into one OpenTelemetry Collector — and what you lose.
LinkMeshOpenTelemetryObservability
Semantic Conventions in Practice
linkmesh.io
Services drift on attribute names until your queries quietly miss data. How to normalize OpenTelemetry semantic conventions at the collector, fleet-wide.