Integrations · OpenTelemetry

Point your Collector at your SIEM. Logs, traces and metrics over OTLP.

If your services already emit OpenTelemetry, you don't need another agent. Add one exporter to the Collector you run today. nano turns the logs into security events you can search and detect on, and keeps traces and metrics next to them.

Try nano open source
How it works

One exporter. No new agent.

nano takes OTLP over HTTPS at /otlp, with an ingest token in the Authorization header. The Collector adds the signal path for you.

# otel-collector.yaml
exporters:
  otlphttp/nano:
    endpoint: https://<ingest-host>/otlp
    encoding: proto
    compression: gzip
    headers:
      Authorization: "Bearer ${env:NANO_TOKEN}"

service:
  pipelines:
    logs:
      receivers: [otlp]
      exporters: [otlphttp/nano]
  1. 01 · Export

    Add nano to the pipeline you have.

    Keep your existing exporters. nano gets a copy of the same data, so your observability tools carry on as before.

  2. 02 · Verify

    Every request is checked.

    nano checks the token on every request, with no cache. A revoked token stops working on the next request. The logs land in the token's source, and that source's parser runs.

  3. 03 · Store

    Each signal goes where it's useful.

    Logs become OCSF events in ClickHouse. Traces and metrics keep their native OTLP shape in their own tables.

Where it lands

What you can do with each signal.

01 · Your services
  • SDKs
  • OTel Collector
  • Fluent Bit
Unchanged Your observability tools keep their copy
02 · /otlp OTLP/HTTP · protobuf · gzip

Token checked on every request

Revoked tokens stop on the next request

03 · nano
  • Logs OCSF events Search, dashboards, detections
  • Traces otel_spans trace, services, Observability
  • Metrics otel_metrics metric, Observability
fig 01 · One exporter in. Each signal lands where it is useful.
SignalWhere it landsWhat you use it for
Logs Routed to the token's source and parsed to OCSF, like any other push event. trace_id and span_id are kept. nPL search, dashboards, detections and alerts
Traces One row per span, in native OTLP shape, in their own ClickHouse table The Observability pages, and the trace and services search commands
Metrics One row per data point, in native OTLP shape, in their own ClickHouse table The Observability pages, and the metric search command

Search your OTLP logs with metadata.log_source="otlp_log". Traces and metrics aren't OCSF events, so detections run on your logs, not on spans or data points. Search commands

Why a SIEM

Your apps are part of the attack surface.

Coverage

See what your services see.

Failed logins, unusual API calls and errors from your own applications are security signals. If they're already in OpenTelemetry, they can be in your SIEM today.

One pipeline

Don't run a second agent.

The Collector your platform team runs for observability can send security a copy. There's one place to change sampling, filtering and routing.

Investigation

Follow the request.

nano keeps the trace_id and span_id that come with your OTLP logs. When a log line looks wrong, look up its trace with the trace command and see what else that request touched.

FAQ

Questions we get asked.

Does nano support OpenTelemetry?

Yes. nano accepts OpenTelemetry logs, traces and metrics over OTLP/HTTP. Logs are parsed into OCSF events that search, dashboards and detections use. Traces and metrics are stored in their own ClickHouse tables in native OTLP shape.

Is OTLP ingest in the free edition?

Yes. OTLP ingest is part of nano's open-source engine (AGPL-3.0) and of nano Self-run. You don't need a license to send OpenTelemetry data to nano.

Which OTLP encoding does nano accept?

Protobuf over HTTP, with gzip allowed. Set your exporter to http/protobuf: the Collector's otlphttp exporter with encoding: proto, which is its default, or OTEL_EXPORTER_OTLP_PROTOCOL=http/protobuf for an SDK. OTLP/JSON isn't supported; the endpoint answers 415 Unsupported Media Type.

Can I send OTLP over gRPC?

On the open-source stack, Vector also listens for raw OTLP on gRPC port 4317 and HTTP port 4318. Those ports don't check a token and are bound to localhost by default, so keep them on a trusted network. Everywhere else, and on nano Self-run, use the authenticated /otlp endpoint over HTTPS.

How is OTLP traffic authenticated?

Each request carries an ingest token as a Bearer header. nano checks the token's current state on every request, with no cache, so a revoked token stops working on the next request. The logs land in the source the token is bound to, and that source's parser runs.

Do detections run on traces and metrics?

No. Detections run on OCSF events, which means your OTLP logs. Traces and metrics aren't OCSF events, so you open them from the Observability pages or with the trace, services and metric search commands.

Can I send OTLP from Fluent Bit or Cribl?

Yes. Fluent Bit's opentelemetry output can post logs to the /otlp/v1/logs path with your token, and the docs cover Cribl Stream as a destination. Anything that speaks OTLP/HTTP with protobuf will work.

Is nano an APM?

No. nano is a SIEM. It keeps traces and metrics so you can follow a request during an investigation, next to the security events around it. Keep your observability tools for performance work, and send nano a copy from the same Collector.

Sources

Read the details.

Start with nano.

From US$149/month at 5 GB/day, US$279/month at 10 GB/day (list price, billed monthly; 10% off billed annually). Every feature in one license, and one license covers two installs.

Buy nano
Talk to sales