Platform · Search
Case Study · Healthtech
Healthcare Platform Testing: Playwright Automation and Artillery Load Tests
Manual testing, Playwright automation and load tests on two layers, from 1,000 to 10,000 virtual users, for an AI platform life sciences teams use.
engagement / ai-healthcareCompleted
- Client
- AI healthcare intelligence platform for life sciences teams, unnamed
- Team
- SprintSynergy QA team, led by our founder
- Engagement
- About 6 months, completed
- Manual Testing
- 200+ test cases
- Automation
- Playwright with JavaScript
- Performance
- Artillery with Playwright (browser load), JMeter (APIs), 1,000 to 10,000 virtual users
200+Manual Test Cases
60%of Manual Test Cases Automated
10KVirtual Users at Peak Load
2Performance Layers Tested: Browser and API
The Result
What the Platform Got
- Releases with no blocker or critical bugs. The platform's features were covered by 200+ manual test cases, 60% of them automated in Playwright.
- Speed checked where users feel it. Browser load tests in Artillery and API load tests in JMeter, from 1,000 to 10,000 virtual users.
Overview
Features and Speed, Tested Together
A SprintSynergy team led by our founder ran QA on an AI-powered healthcare intelligence platform that life sciences teams use to find and work with medical experts. We wrote 200+ manual test cases and built an automation framework in Playwright with JavaScript, which automated 60% of them. Performance testing ran on two layers, from 1,000 to 10,000 virtual users. Artillery drove real browsers through Playwright, and JMeter put the APIs under load.
Load Tests · Two Layers
BrowserArtillery · Playwright
Real browser sessions under load, so the results show what users see.
APIJMeter
The platform's APIs under the same range of load.
1,000 Virtual Users10,000 Virtual Users
The Problem
Correct, and Fast Under Load
An AI platform for life sciences teams has two jobs at once. Its features have to work correctly, and it has to stay fast when many people use it together. A slow page makes users doubt the data behind it.
Speed has to be checked at two levels. The APIs can answer quickly while the page in the browser is still slow, because of what happens after the data arrives.
What We Tested
Four Angles on One Platform
- Features: the platform's functions, through 200+ manual test cases.
- Automated Checks: 60% of the manual test cases, automated in Playwright with JavaScript.
- Front-End Performance: real browser sessions under load, from 1,000 to 10,000 virtual users, through Artillery and Playwright.
- API Performance: the platform's APIs under the same range of load, in JMeter.
How We Built It
Playwright, Artillery and JMeter
- 200+ manual test cases for the platform.
- A Playwright automation framework in JavaScript, covering 60% of the manual test cases.
- Browser load tests: Artillery runs Playwright scripts in real browsers, so the results show what users experience, not only how fast the server answers.
- API load tests in JMeter.
- Load stepped from 1,000 to 10,000 virtual users on both layers.
What We Found
Three Bugs Testing Caught
Platform · Doctor Details
Doctor Details Did Not Display Correctly
Platform · Graphs
Graphs Did Not Load
What We Delivered
Handed Over
- 200+ manual test cases for the platform.
- A Playwright framework in JavaScript, with 60% of the manual test cases automated.
- Browser load tests with Artillery and Playwright, up to 10,000 virtual users.
- API performance tests in JMeter, up to 10,000 virtual users.
Questions
Frequently Asked Questions
Yes. This platform's framework was written in JavaScript. Most of our current work, such as our e-commerce projects, uses TypeScript.
Need to Know Your Platform Holds Up Under Load?
Book a free 30-minute QA strategy call. A written QA plan follows within 48 hours.
Book a Free QA Strategy Call