robot-triage CLI auto-flags failures in ROS bags
No more scrubbing hours of bag data; this CLI does it for you.
Manually scanning hours of ROS bag data for failures is tedious and error-prone, especially when robots generate large logs with only a few critical moments of interest. robot-triage, a new open-source CLI created by v2pir, automates this first pass. It ingests .bag or .mcap files and automatically flags six common failure types: sensor dropouts, error bursts, pose jumps, divergence between commanded and actual states, frozen values, and out-of-range readings. The tool requires zero configuration—it learns each topic's normal rhythm directly from the bag file itself, adapting to any robot. This makes it usable out of the box for both ROS 1 and ROS 2 users.
For each detected anomaly, robot-triage trims a small MCAP clip and generates a Foxglove link pointing to that exact moment in time, enabling instant visual inspection. Advanced users can optionally provide a YAML file to tune detection thresholds for specific topics. The tool is still new, and its creator encourages the community to test it on real-world bags, reporting both false positives and missed failures. By eliminating the need to manually scrub through terabytes of log data, robot-triage significantly accelerates robot debugging, letting engineers focus on root causes rather than data hunting.
- Detects 6 common failure types: sensor dropouts, error bursts, pose jumps, cmd vs actual divergence, frozen values, and out-of-range readings.
- Generates small MCAP clips with Foxglove links for instant visual inspection of flagged moments.
- Zero configuration required; tool auto-learns topic patterns from the bag file itself.
Why It Matters
Robotics engineers can now debug bag files in minutes instead of hours, speeding up failure analysis.