Cloud load testingHold up to 1,000 virtual users on your site, from the cloud.
A cloud load test holds up to 1,000 virtual users on one verified site from cloud infrastructure, so you see how it answers at the level a serious launch might bring, before the launch does.
50 cloud load tests a month on Business.
3,175
requests a second
1,000
virtual users held
419 ms
p95 response time
From the sample report: 402,389 requests, none failed.
The highlights.
A light test tells you the site is quick. It cannot tell you what happens when a thousand people arrive together. How the site answers at that level is the thing to know before a big day.
Four steps. From cloud to verdict.
Verify the asset
Every test runs against a domain you've proven you own. No testing systems you don't control.
Start the cloud run
A dedicated cloud machine is started for your run and generates genuine concurrent load at a scale a single edge worker can't reach.
Hold the level
The number of users you chose is held for the whole run while we record response time, throughput, and errors.
Read the verdict
A risk rating set by fixed rules, and an AI summary of what the numbers mean and what to check.
A level held, and where the time went.
The report gives the users held, throughput and p95, the error rate, and where the time went. In the sample, time to first byte was most of the total: the cost was the backend, not the network.
Cloud load report
Sample reportExample site
Level held
1,000 users
Throughput
3,175 req/s
Error rate
0%
p95 response
419 ms
AI summary
A genuinely strong result: 1,000 concurrent users from real cloud infrastructure, 402,389 requests, zero errors, and p95 holding at 419 ms with almost no tail. The detail worth acting on is where the time goes — TTFB averages 255 ms of the 293 ms total, so backend processing, not network transfer, is the cost. One 12.7-second outlier in the max is worth tracing to a cold start or a slow query.
A performance assessment, not a penetration test.
Branded, downloadable, shareable: the PDF carries the same page, ready for a review or a vendor questionnaire.
What a cloud load test measures.
Four figures from the sample run, and why each one matters.
419 ms
p95 latency
95% of requests finished faster than this. The honest measure of user experience — averages hide the slow tail that frustrates real people.
255 ms
TTFB
Time to first byte — how long the server thinks before it answers. When this dominates total time, the cost is backend processing, not the network.
3,175 req/s
Throughput
Requests per second over the run. What your stack answered while this many users were held on it — a reading at one level, not a ceiling.
0.00 %
Error rate
Share of requests that failed under load. Zero across 402K requests is the signal that nothing buckled when the pressure was on.
What you get.
- Up to 50 cloud load tests every month on Business
- 50 to 1,000 concurrent virtual users from real cloud infrastructure
- How your site answers at a level you choose, held for up to two minutes
- Response time, throughput, and error rate at scale
- An AI summary of what the numbers mean and what to check
- See all features ›
Edge load test vs. Cloud load test
Both drive traffic at a verified asset — the difference is scale and where it runs. Pick the lighter edge test for a quick read, the cloud test to see the site hold a level you choose.
| What is compared | Edge load test (Pro) | Cloud load test (Business) |
|---|---|---|
| Where it runs | Cloudflare's edge network | Isolated cloud infrastructure |
| Concurrency | 1 to 6 connections | 50 to 1,000 virtual users |
| Best for | A fast, low-cost sanity check | Holding a level you choose before a launch |
| Engine | Lightweight edge workers | A load-testing engine on dedicated infrastructure |
Is it the right tool?
A strong fit when
- A launch, a sale or a press moment is coming and you want to see the site at that level first.
- You are deciding how much infrastructure to pay for.
- The Pro load test came back clean and you want a far heavier one.
Not the best fit when
- You want a quick check before a routine release. The Pro load test is lighter.
- You need the load held for a long time. Cloud stress testing does that.
- You need a penetration test: a person trying to break in. We do not offer one.
A cloud load test sends heavy, real traffic. It runs only against a site you have verified, you choose the users and the duration, and a queued or running test can be cancelled from the Jobs page.
Included in Business.
Business
50 cloud load tests a month, up to 1,000 virtual users, 100 sites
$199 a month₹19,499 a month in India
In more detail.
What a cloud load test measures
A cloud load test holds 50 to 1,000 concurrent virtual users on one domain you have verified you own and measures the same honest signals as the edge test — response time across percentiles, time to first byte, throughput, and error rate — at a scale closer to a real launch or campaign. Each run holds one level, so the output is how your site answered at that level; run it again at a higher one to see where latency climbs or errors begin. Those are the numbers you want before a big day, not during it.
How it works, and why cloud rather than edge
Instead of a single edge worker, a cloud load test runs from cloud infrastructure, generating genuine concurrent load: the number of virtual users you choose, from 50 to 1,000, is held for 20 to 120 seconds while performance is recorded. It uses a dedicated load-testing engine and is available on the Business plan. Every run is gated behind verified domain ownership.
When to run one, and what's in the report
Reach for a cloud load test before a product launch, a sale, a press moment, or any event where you expect real cloud-scale traffic. The report is a downloadable, branded PDF: the level held, throughput and p95, the error rate, a risk rating, and an AI summary that points to what to check first. Example result: 1,000 concurrent users sustained across 402,389 requests with zero errors and a p95 of 419 ms; the detail worth acting on was where the time went — time to first byte was most of the total, which points at backend processing rather than the network. If your site sits behind a CDN, a strong result may partly reflect the CDN rather than your own server — keep that in mind when reading the numbers.
What a cloud load test is not
A cloud load test is a performance assessment, not a penetration test: it measures how your site behaves under generated traffic; it does not probe application logic or attempt exploitation. A penetration test is a separate, human-led engagement. BCDRPulse does not offer one. It generates real, heavy load, so you run it only against a domain you have verified you own, and you set the parameters of the run.
Questions people ask first.
Is this a penetration test?
No. Cloud load testing is an automated performance assessment — it measures how your site behaves under generated traffic. It does not probe application logic or attempt exploitation. A penetration test is a separate, human-led engagement. BCDRPulse does not offer one.
Will it take my site down?
It generates real load, so you should only run it against an asset you own and control — which is why every test is gated behind verified ownership. You choose the concurrency and duration, and a queued or running test can be cancelled from the Jobs page.
Does it work behind a CDN like Cloudflare?
Yes. Where a CDN serves cached responses, the results reflect the edge rather than your own server — keep that in mind when reading the numbers.
How many cloud load tests can I run?
Up to 50 every month on the Business plan, against any of your verified assets. The limit resets at the start of each calendar month.
How is the cloud load test different from the edge load test?
The edge test is a quick, light sanity check from Cloudflare's network. The cloud test holds far more concurrent users, up to 1,000, from cloud infrastructure — the difference is scale and where it runs.
How many concurrent users can it simulate?
From 50 to 1,000 concurrent virtual users, in steps of 50 — well beyond the six the edge test reaches.
Does it find my 'capacity ceiling'?
Not in one run. A capacity ceiling is the point where your stack stops keeping up: adding more concurrent users no longer raises throughput, latency climbs sharply, or errors start. Each run holds one level, so you find it by running again at higher levels and comparing the reports.
Know how it holds before the big day.
Signing up is free and starts uptime monitoring for 10 sites. Running cloud load testing needs the Business plan.
- 1,000 users, 3,175 requests a second and a 419 ms p95 are the figures in the sample report, not a customer's site.
- Allowances are per calendar month and can be used on any of your verified sites.
- Introductory prices, before GST or VAT. Regular prices are $99 and $399 a month (₹9,699 and ₹38,999 in India).