Loading system evidence
Loading system evidence
Operate ROS 2 systems through network failures, container boundaries, deployments, rollback, and fleet incidents.
Sign in to start - freeOutcome
Build
A monitored two-site fleet deployment with release gates, rollback, and an incident-response evidence bundle.
Sample incident
Restore discovery without moving the fleet onto the default domain.
Full syllabus
Trace participant discovery, domain isolation, middleware selection, and endpoint compatibility.
Inspect multicast, routes, firewalls, VLANs, VPNs, and unstable wireless links.
Package ROS applications while preserving users, devices, volumes, graphics, and DDS network behavior.
Define service identities, dependencies, readiness, restart limits, and safe shutdown.
Design actionable health checks, structured logs, service metrics, and causal traces.
Version images, configuration, database migrations, environment compatibility, and rollout cohorts.
Design reversible software/configuration changes and collect bounded, privacy-aware incident evidence.
Triage, contain, reconstruct, repair, verify, and communicate multi-robot incidents.
Two safety-gating bugs from a working warehouse robot, the kind that hurt someone on real hardware: an override that bypasses E-STOP, and a directional gate wired backwards.
Playgrounds
Small, functional experiments that reveal the math and evidence behind this course’s failures. Open them alongside the labs.
Step a managed (lifecycle) node through its states and see which transitions are legal from unconfigured, inactive, and active.
lifecycle node / managed node / state machine
Live controlsCompute publish bandwidth from rate, message size, and subscriber count, and see when a link saturates and starts shedding messages.
topic bandwidth / DDS / message rate
Live controlsSee how ROS_DOMAIN_ID and ROS_LOCALHOST_ONLY decide whether two nodes ever discover each other across hosts.
ROS_DOMAIN_ID / DDS discovery / ROS_LOCALHOST_ONLY