Four problems. One control plane.

Problem

OpenTelemetry Collectors are great per-host, but managing dozens or hundreds across a fleet means manual config drift, version skew, and SSH-driven ops.

LinkMesh

Enroll every Collector with the LinkMesh control plane over a single port-443 connection — otelcol-contrib via OpAMP, Grafana Alloy via remotecfg. The fleet UI shows status, version, throughput, and config state for every node. Push config changes from one place; collectors apply them with hot-reload where possible, restart where needed.

Problem

Sensitive fields — credit cards, emails, phone numbers, internal IDs — slip into logs all the time. They reach observability backends where they're indexed, retained, and eventually queried by people who shouldn't see them.

LinkMesh

Apply masking processors at the Collector level, BEFORE telemetry leaves the host. Built-in templates for credit cards, emails, phone numbers; custom regex/OTTL rules for your specific fields. Masking runs client-side in customer infrastructure — sensitive values never reach the network.

Problem

Most telemetry is noise. Health checks, polling, debug logs, retry storms — they bloat your retention bill and bury the signals you actually need.

LinkMesh

Drop, sample, aggregate, and rate-limit at the source. Filter rules that read clean (drop where service.name = "health-check") instead of OTTL spaghetti. Per-component throughput metering at every stage — receivers, processors, exporters — so you can see exactly where volume is reduced.

Problem

A single telemetry pipeline can't serve every team. Security wants logs in Splunk; SRE wants metrics in Prometheus; finance wants slow queries in BigQuery.

LinkMesh

Define routes by label, environment, tenant, or service. The same source can fan out to multiple destinations with different processing chains for each. Every route, pipeline, and destination edit lands as a Git commit in the GitOps config repo (ADR 0013) — full diff + authorship for every change.