AI Scenarios
The AI can propose stress scenarios for your design. It analyzes your architecture, picks a realistic failure mode, and configures the deterministic runtime to execute it.
How it works
Section titled “How it works”- Click Simulate in the header or ask the AI chat to stress test your design
- AI analyzes your design and proposes a scenario (load spike, component outage, cascading delay)
- The deterministic runtime executes the scenario step by step
- Results show a timeline with health percentage, first failure point, cascade path, and fix recommendations
- Request playback auto-starts so you can see the flow on the canvas
What AI controls vs. what it doesn’t
Section titled “What AI controls vs. what it doesn’t”| AI proposes | Runtime computes |
|---|---|
| Scenario parameters (load, outage target, delays) | Actual request routing and queue behavior |
| Narrative framing | Health percentage per step |
| Fix recommendations | First failure point and cascade path |
The AI never fabricates outcomes. It sets the scenario, the engine runs it, and the results are deterministic.
Manual stress testing
Section titled “Manual stress testing”You can also run stress tests manually without AI using the Stress Test tab in the Runtime Panel (R key). This gives you:
- Quick scenario presets (Baseline, Peak Traffic, Outage, Slowdown)
- Traffic controls (seed packets 1-20, traffic multiplier 1-5x, arrival pattern)
- Fault injection (kill or slow down components, add latency)
See Stress Testing for full details on manual controls.