[RUM-12600] Continuous Benchmarking

[amortemousque]

Motivation

We want to monitor the impact of performance-intensive features like Replay and Profiling.
This PR introduces benchmark tests to measure their effect.

Measurements

• Memory, CPU using the Chrome DevTools Protocol. Same implementation as /performances, except it only collects the total consumption.
• Network using the Playwrite page.on('requestfinished') API
• Web Vitals using Google’s web-vitals package

Scenarios

Run two scenarios on two applications:

• React Heavy SPA: Emulating a observability app cpu intensive, Web Vitals in the "Good" score range.
• React Shopist Like: Lightweight E-Commerce App, targeting Web Vitals in the "Needs improvement" score range.

This scenario runs under four different configurations:

• none: no SDK
• none_with_header: SDK disabled but profiling headers present
• rum: RUM SDK only
• rum_replay: RUM with Session Replay
• rum_profiling: RUM with Profiling

Scheduling

A new performance-benchmark GitLab job runs every 30 minutes.
Each run executes the tests 15 times.

Visualization

Local: displays a summary table in the console

Datadog dashboard

Changes

• add test/app/react-heavy-spa: app to benchmark test runs
• add test/performance: Playwright performance benchmarks
• move e2e/lib/framework/httpServer to test/lib for reuse in perf tests
• add performance-benchmark CI job

I haven’t replaced /performances yet, but if we agree on this approach, I’ll remove it in a follow-up PR.

Test instructions

yarn test:performance

Checklist

• Tested locally
• Tested on staging
• Added unit tests for this change.
• Added e2e/integration tests for this change.

[datadog-official]

✅ Tests

🎉 All green!

❄️ No new flaky tests detected
🧪 All tests passed

🎯 Code Coverage
• Patch Coverage: 100.00%
• Total Coverage: 92.67% (+0.00%)

View detailed report

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: 9bac9cf | Docs | Datadog PR Page | Was this helpful? Give us feedback!}

[cit-pr-commenter]

Bundles Sizes Evolution

📦 Bundle Name	Base Size	Local Size	𝚫	𝚫%	Status
Rum	164.05 KiB	164.05 KiB	0 B	0.00%	✅
Rum Profiler	4.84 KiB	4.84 KiB	0 B	0.00%	✅
Rum Recorder	19.50 KiB	19.50 KiB	0 B	0.00%	✅
Logs	55.91 KiB	55.91 KiB	0 B	0.00%	✅
Flagging	944 B	944 B	0 B	0.00%	✅
Rum Slim	121.33 KiB	121.33 KiB	0 B	0.00%	✅
Worker	23.63 KiB	23.63 KiB	0 B	0.00%	✅

🚀 CPU Performance

Action Name	Base CPU Time (ms)	Local CPU Time (ms)	𝚫%
RUM - add global context	0.004	0.0046	+15.00%
RUM - add action	0.0114	0.0128	+12.28%
RUM - add error	0.0112	0.0116	+3.57%
RUM - add timing	0.0026	0.0027	+3.85%
RUM - start view	0.0031	0.0032	+3.23%
RUM - start/stop session replay recording	0.0007	0.0007	0.00%
Logs - log message	0.0132	0.0143	+8.33%

🧠 Memory Performance

Action Name	Base Memory Consumption	Local Memory Consumption	𝚫
RUM - add global context	25.77 KiB	25.80 KiB	+32 B
RUM - add action	93.84 KiB	112.78 KiB	+18.94 KiB
RUM - add timing	24.90 KiB	24.33 KiB	-588 B
RUM - add error	98.13 KiB	116.38 KiB	+18.25 KiB
RUM - start/stop session replay recording	24.66 KiB	23.83 KiB	-849 B
RUM - start view	448.19 KiB	487.34 KiB	+39.15 KiB
Logs - log message	42.73 KiB	43.23 KiB	+513 B

🔗 RealWorld

[BenoitZugmeyer]

💭 thought: ‏I think there are more idiomatic ways to declare common test logic in playwright using fixtures. But personally I don't mind your approach, as I find it more explicit than using playwright magic.

[rtrieu]

approved with some very minor suggestions!

[rtrieu]

Suggested change

	Customize RUM configuration via URL params:
	Customize RUM configuration using the following URL params:

[rtrieu]

Suggested change

	A complex, JavaScript-heavy single-page application built with React + TypeScript for performance benchmarking of Datadog Browser SDK features.
	This is a complex, JavaScript-heavy single-page application built with React + TypeScript for performance benchmarking the Datadog Browser SDK.