Problems / Multi-agent and harness, hard
A deep-research system fans out to 5 subagents that each browse the web; the browser pool has 16 browsers.
Free to open. Runs in your browser, no sign-up.
Success ≥ 95% of tasks, p99 ≤ 90 s.
2 requests a second for 45 seconds: 100% task (writes).
What each component does in this design, and what calls it. The simulator mimics each one from these settings.
| Component | Its job here | Settings |
|---|---|---|
| Load balancerLoad balancer | Traffic enters here.Calls Agent API.Layer 4 balancers route by IP and port; layer 7 balancers read HTTP and can route by path or header. | – |
| Agent APIApp service | Load balancer forwards to it.Calls Agent orchestrator.Little's law: throughput ≈ concurrency ÷ latency. 200 slots at 50 ms serve about 4,000 req/s. | 3 replicas × 500 slots · timeout 180 s |
| Agent orchestratorAgent orchestrator | Agent API calls it.Calls Subagent pool.An agent loops: think, call a tool, read the result, repeat until done. | 400 slots · 1 steps |
| Subagent poolSubagent pool | Agent orchestrator calls it.Calls Browser tool.Multi-agent research systems fan a task out to parallel subagents and use many times more tokens than a single agent. | 400 slots · 5 steps |
| Browser toolBrowser tool | Subagent pool calls it.Browser tools let agents read and act on web pages, which is slow and brittle. | 16 slots |
Try it first. Open one hint at a time if you get stuck.
Browsing is slow: seconds per page.
Multiply by the fan-out and the number of turns to get the real call rate on shared models and tools.
Add browser capacity (or reduce fan-out).