The OpenTelemetry-native
infrastructure agent.

Collect host and infrastructure metrics as OpenTelemetry — push via OTLP to Prometheus, VictoriaMetrics or Grafana, or serve PRTG and Nagios natively. Your stack, your choice.


SenHub Agent
one collector running on every host
# collect host & infrastructure metrics
cpu · memory · disk · network · logs · databases · Citrix · NetScaler · Veeam …
▼   one OpenTelemetry-native pipeline   ▼
OTLP Prometheus PRTG Nagios

One OpenTelemetry pipeline, every output. Push OTLP, scrape Prometheus, or serve PRTG and Nagios — all first-class, all from the same agent. No per-tool agents to deploy.


Why SenHub Agent

Observability has standardized on OpenTelemetry — everywhere except the infrastructure edge, where each tool still ships its own agent. SenHub Agent closes that gap: a single OTel-native collector for hosts, databases and infrastructure. Emit OTLP into a modern stack, or serve PRTG, Nagios and Prometheus natively — every output is first-class and fully supported. Adopt OpenTelemetry at your own pace, without leaving the tools you rely on.


What it does

OpenTelemetry-native

Every metric is modelled on OpenTelemetry semantic conventions from the start, then emitted straight out as OTLP — no proprietary protocol, no translation layer, no lock-in.

Plugs into what you run

OTLP and Prometheus feed a modern stack — VictoriaMetrics, Grafana, any OTel backend. PRTG, Nagios and Prometheus are first-class outputs too — natively supported, your choice.

Open-core & lightweight

Host self-observability and universal collection are free and Apache-2.0; deep vendor probes are the paid layer. One ~9 MB binary, no runtime, no dependencies.


Light footprint, runs anywhere

One static Go binary for Linux, Windows and macOS (amd64 and arm64) — no JVM, no Python, no agent framework. Drop it onto bare metal, a VM, a container or a Kubernetes node: it collects locally and pushes OTLP out, replacing a stack of per-tool agents with a single low-overhead process.

1
static binary
≈9 MB
compressed download
0
runtimes or dependencies
5
OS / arch targets
Linux Windows macOS · amd64 arm64 · Docker Kubernetes · soon

Quick start

Three steps, no agent-per-tool sprawl. The example is Linux; Windows and macOS follow the same flow with their own archive.

1

Install the binary as a service

$ unzip senhub-agent-linux-amd64.zip -d /opt/senhub/bin/
$ /opt/senhub/bin/senhub-agent install
$ /opt/senhub/bin/senhub-agent start
2

Declare probes and an output in YAML

# agent-config.yaml
probes:
  - { name: cpu,         type: cpu,         params: { interval: 30 } }
  - { name: memory,      type: memory,      params: { interval: 30 } }
  - { name: logicaldisk, type: logicaldisk, params: { interval: 30 } }

storage:
  # push OpenTelemetry to your collector / backend
  - name: otlp
    params: { protocol: grpc, endpoint: "otel-collector:4317" }
  # …and/or expose pull endpoints (Prometheus, PRTG, Nagios)
  - name: http
    params: { port: 8080, endpoints: ["prometheus", "prtg", "nagios"] }
3

It's already shipping

# OTLP is already pushing — scrape Prometheus or check from Nagios too
$ curl -s localhost:8080/api/<agentkey>/prometheus/metrics
$ curl -s localhost:8080/api/<agentkey>/prtg/metrics/cpu

Validate any config with agent config check and inspect the live pipeline with agent status --otlp. Pro probes unlock via agent license activate <token>. Full reference at the documentation.


What it monitors

From host self-observability to deep vendor infrastructure — one agent, one config format, OpenTelemetry-native throughout.

Host & OS Free

CPU, memory, logical disk, network interfaces.

Host logs Free

Linux journald shipping (linux_logs).

Databases Pro

MySQL, PostgreSQL — OTel-first metric mappings.

Virtualization / VDI Pro

Citrix Virtual Apps & Desktops sessions and machines.

Application delivery Pro

Citrix NetScaler ADC — vServers, services, load balancing.

Server hardware Pro

Redfish — iDRAC, iLO and other BMCs.

Data protection Pro

Veeam Backup & Replication jobs and sessions.

Synthetic & wireless Pro

Gateway ping, web-app reachability and load, Wi-Fi signal.

Syslog & events Pro

Syslog relay (RFC 3164 / 5424) and custom HTTP events.

Free probes ship in the Apache-2.0 build and need no licence. The full, growing catalogue lives at the documentation.


Integrations

One neutral, OpenTelemetry-mapped data store; each output is a serialization of the same metrics — pull or push, your choice.

OutputModeFormatStatus
OpenTelemetryPushOTLP / gRPC Stable
PrometheusPull/metrics (text 0.0.4) Stable
PRTGPullNative PRTG JSON Stable
NagiosPullPlugin text + perfdata Stable

OTLP and Prometheus drop straight into any OpenTelemetry backend — VictoriaMetrics, Grafana, and the rest. PRTG and Nagios are first-class, fully-supported outputs — use whichever fits your team.


Open-core

The free tier is enough to monitor a whole fleet. Paid tiers add vendor and infrastructure probes — and the licence validates fully offline, with no call home.

TierWhat you getActivation
Free Host self-observability — CPU, memory, disk, network, journald logs. Unlimited hosts. None — Apache-2.0 build
Pro Vendor & infra probes — databases, Citrix, NetScaler, Redfish, Veeam, syslog, synthetic checks. agent license activate
Enterprise Every probe, including future ones (wildcard) — e.g. IBM i. agent license activate

Licences are signed JWTs verified locally — offline by design, with a 7-day grace period after expiry.


Part of SenHub

SenHub Agent is open source and built by the team behind SenHub (Sensor Factory). It stays deliberately vendor-neutral — it speaks open standards and feeds whatever you already run, with no lock-in and no mandatory backend.

New here? The full install guide, probe catalogue and configuration reference live at the documentation.


Follow along

SenHub Agent is built in the open. The fastest way to support it is to star the repository — it tells us the direction resonates. Read the docs to get running in minutes, or open an issue to shape what's next.