Advanced
Everything here assumes you already have nodes running and are now deciding how they run: which execution mode the scheduler uses, what happens when one node misbehaves, and what you need to change before the same code drives real hardware.
Nothing on this page is required to build a working application. Start with Core Concepts and the Quick Start; come back when you hit a question one of the pages below answers.
Execution and scheduling
How work is ordered, distributed, and made reproducible.
| Page | What it covers |
|---|---|
| Execution Modes | Sequential and Parallel scheduler execution modes, and when each applies |
| Scheduler Configuration | The fluent node builder API and how execution is configured |
| Scheduling Intelligence | Execution class annotations that inform scheduling |
| Deterministic Execution | Reproducible execution for testing, debugging, and analysis |
| Real-Time Configuration (RtConfig) | System-level real-time settings for deterministic execution |
Fault tolerance and safety
What the runtime does when something goes wrong.
| Page | What it covers |
|---|---|
| Circuit Breaker | Preventing one failing node from cascading through the system |
| Safety Monitor | Watchdogs, WCET enforcement, and emergency handling |
Observability and debugging
Recording what happened so you can inspect it afterwards.
| Page | What it covers |
|---|---|
| BlackBox Flight Recorder | Event recording and post-mortem debugging |
| Record & Replay | Tick-perfect replay for debugging, testing, and analysis |
Transport and deployment
| Page | What it covers |
|---|---|
| Network Backends | Shared-memory IPC and the other communication backends |
| Production Deployment | Deploying to production robots with real-time features enabled |
Related
- Development — the tooling around your code: testing, parameters, the CLI, and static analysis
- Performance — measuring and tuning, once the behaviour above is correct