Introduction
TraceHouse™ is the observability platform for ROS2 engineers. It gives you distributed tracing, fleet health monitoring, and telemetry analysis — all queryable with ROSQL, the open source query language built for ROS2.
How it works
Three components work together to collect and surface robot telemetry:
| Component | Role |
|---|---|
| RobotOps Trace SDK | Small libraries you add to your nodes (C++ + Python) that emit deterministic, in-process OpenTelemetry traces — auto-instrumenting every callback and action |
| TraceHouse agent | On-robot "carrier" binary that receives spans from the SDK over a local socket and streams them to TraceHouse — and also collects logs, metrics, TF, and MCAP |
| ROSQL | SQL-like query language for exploring traces, metrics, and logs across your fleet |
:::note rmw_robotops is deprecated
The legacy passive rmw_robotops RMW shim has been superseded by the SDK above.
See Related Libraries.
:::
What's in these docs
- Quickstart — Install the SDK + agent and start tracing in minutes
- Instrumentation — Auto-init, integrations, annotations, and what each span carries
- System Requirements — Supported OS, ROS2 distros, and architectures
- TraceHouse Agent Overview — What the agent (carrier) collects and how it works
- Glossary — Key terms across the TraceHouse ecosystem
- FAQs — Common questions
- Related Libraries — the Trace SDK repos, rosql, robotops_msgs, robotops_config