Overload Behavior Simulator

See what happens when your system hits its limits

Watch latency spike, queues build, and requests get dropped — then switch strategies and see the tradeoff instantly.

Queues look fine until they don't

When a system hits capacity, adding a queue feels like the responsible choice. Fewer errors. Calmer dashboards. But under sustained load, that queue is silently converting rejected requests into latency — and the longer it gets, the worse every user's experience becomes.

By the time you notice, your p95 is through the roof, timeouts are cascading, and the queue that was supposed to help is the reason your system feels broken. The overload was always there. The queue just hid it.

LoadLens makes this tradeoff visible instantly.

⬤ Fail-Fast

Completed290
Rejected300
P95 latency1.0s
Max waiting0

⬤ Bounded Queue

Completed290
Rejected300
P95 latency1.8s
Max waiting5
Same throughput. Same rejections. 80% worse latency.

Built for understanding, not guessing

Instant simulation

Powered by async-bulkhead-ts — the same admission control library used in production systems. Drag a slider, see the result immediately.

Side-by-side comparison

Fail-fast and bounded queue run simultaneously. Same parameters, different strategies. The difference is impossible to miss.

📊

Four core charts

P95 latency, throughput, waiting requests, and rejections over time. Each one tells part of the overload story.

🎯

Guided walkthrough

Step through overload behavior: from healthy → saturation → collapse. See exactly where things go wrong.

🧠

Dynamic insights

LoadLens surfaces the core tradeoff automatically: how many rejections the queue absorbed, and what it cost in latency.

📤

Export & present

Save scenarios locally. Export charts as PNG. Use LoadLens in architecture reviews, postmortems, and conference talks.

Common questions

Is this a production monitoring tool?

No. LoadLens builds intuition about overload behavior through simulation. It's an educational tool, not a replacement for observability, capacity planning, or load testing. Think of it as a whiteboard that runs real admission control logic.

How accurate is the simulation?

LoadLens uses async-bulkhead-ts — a real concurrency-limiting library — under the hood. The simulation matches real bulkhead behavior for the parameters you configure. It uses deterministic arrivals and fixed work durations, which makes it ideal for building intuition. Real systems add variance, retries, and network effects that LoadLens intentionally omits for clarity.

Is LoadLens open source?

The underlying concepts and simulation model are transparent, but the full product is not open source. LoadLens is designed to be used directly — no setup required.

Does LoadLens collect any data?

No. LoadLens is a static web app. There is no backend, no analytics, no telemetry. Your simulations run entirely in your browser. Nothing leaves your machine.

Can my team use this?

Yes — LoadLens is a static web app. Share the URL with anyone on your team. No accounts, no seat limits.

What if I want to plug in real traffic data?

That's intentionally out of scope. LoadLens is sharpest when it stays simple: a few controls, clear output, immediate insight. If you need production traffic simulation, you want a load testing tool. LoadLens helps you understand why your system behaves the way it does under load — before you run those tests.