I built a tool to find out if your site survives a traffic spike before it actually happens

After many years working in infrastructure, SRE and platform engineering, I kept seeing the same story: a site handles normal traffic just fine, then goes down on the one day that matters most. Black Friday, a post that went viral, a push notification everyone opens at once, a launch during prime time.

Traditional load testing usually measures steady load. But what actually takes sites down is the **shape** of the spike: a near-vertical ramp in a few seconds that catches queues, databases and caches off guard.

So I built **spikeloader** (https://spikeloader.com). It does one thing and tries to do it well: simulate audience spikes.

**How it works:**

– Pick a ready-made scenario (viral post, Black Friday, push notification) or design your own: baseline, ramp, peak and recovery

– Traffic comes from our infrastructure, so there's nothing to install

– Live charts while the test runs: planned vs. served req/s, p50/p95/p99, errors broken down by cause (timeouts, refused connections, 5xx)

– At the end you get a plain-language verdict, like "started degrading at 1:45 and recovered 20s after the peak". Handy for explaining results to people outside infra

**On safety**, which was my biggest concern (a tool that generates crowds can't become a weapon):

– It only tests verified domains (DNS record or file), and verification is re-checked on every run

– Private and internal networks are blocked, even when hidden behind public DNS

– Traffic comes from a fixed IP, so you can allowlist it in Cloudflare/your WAF and the test isn't treated as an attack

– A cancel button stops the test immediately

**There's a free plan, no card required:** up to 500 rps, 10k requests/month and tests up to 10 minutes. Enough to run a real test before deciding anything.

One heads-up: all traffic currently originates from São Paulo, Brazil, so if your users are elsewhere, expect some extra baseline latency in the numbers.

I'd really love feedback from people working in infra, e-commerce or product:

– What kind of spike has taken your system down before?

– What metric or info would you miss in the results?

Criticism welcome, including the harsh kind. 🙂