[Profiler] Remove sample-count threshold from profiler discard logic

[thomasbertet]

Motivation

Short-lived non-SPA navigations (full-page loads with moderate scripting) were being silently discarded by the profiler because they accumulated fewer than 50 stack-trace samples (~500ms of scripting activity), even when the page was active long enough to meet the duration threshold. This caused an empty "No profiles" state in the Profiling UI for use cases like shopist.io where users click quickly across pages.

Additionally, filtering on sample count introduced a selection bias in aggregations: only CPU-heavy pages were retained, skewing flame graphs away from scripting-light but potentially slow navigations.

The trade-off here is collecting more profiles for a slightly better coverage of short-lived navigations.
Nonetheless, we keep discarding profiles <5s if they don't have Long Tasks in them.

Changes

• Removed the minNumberOfSamples check from the profile discard logic in profiler.ts. Profiles are now kept or discarded based solely on duration (minProfileDurationMs) and long-task presence.
• Removed the minNumberOfSamples field from RUMProfilerConfiguration and DEFAULT_RUM_PROFILER_CONFIGURATION.
• Added unit tests covering the three discard cases: below duration threshold with no long tasks (discarded), below duration threshold with a long task (sent), and above duration threshold (sent).

Test instructions

1. Load a non-SPA page with the RUM SDK and profiler enabled.
2. Navigate away quickly (within a few seconds) before reaching the 5s duration threshold.
3. Verify a profile is sent if a long task was captured, discarded otherwise.
4. Run yarn test:unit --spec packages/rum/src/domain/profiling/profiler.spec.ts — 20 tests should pass.

Checklist

• Tested locally
• Tested on staging
• Added unit tests for this change.
• Added e2e/integration tests for this change.
• Updated documentation and/or relevant AGENTS.md file

[cit-pr-commenter-54b7da]

Bundles Sizes Evolution

📦 Bundle Name	Base Size	Local Size	𝚫	𝚫%	Status
Rum	179.65 KiB	179.65 KiB	0 B	0.00%	✅
Rum Profiler	6.17 KiB	6.04 KiB	-132 B	-2.09%	✅
Rum Recorder	27.03 KiB	27.03 KiB	0 B	0.00%	✅
Logs	56.78 KiB	56.78 KiB	0 B	0.00%	✅
Rum Slim	135.50 KiB	135.50 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	0.0039	0.004	+2.56%
RUM - add action	0.0131	0.0137	+4.58%
RUM - add error	0.0117	0.0123	+5.13%
RUM - add timing	0.0025	0.003	+20.00%
RUM - start view	0.0119	0.0119	0.00%
RUM - start/stop session replay recording	0.0006	0.0007	+16.67%
Logs - log message	0.0138	0.0194	+40.58%

🧠 Memory Performance

Action Name	Base Memory Consumption	Local Memory Consumption	𝚫
RUM - add global context	31.78 KiB	32.57 KiB	+810 B
RUM - add action	56.94 KiB	56.94 KiB	-9 B
RUM - add timing	32.72 KiB	33.40 KiB	+694 B
RUM - add error	60.80 KiB	61.40 KiB	+618 B
RUM - start/stop session replay recording	32.10 KiB	31.94 KiB	-168 B
RUM - start view	484.29 KiB	483.78 KiB	-523 B
Logs - log message	95.66 KiB	103.74 KiB	+8.08 KiB

🔗 RealWorld

[datadog-prod-us1-6]

🎉 All green!

❄️ No new flaky tests detected
🧪 All tests passed

🎯 Code Coverage (details)
• Patch Coverage: 0.00%
• Overall Coverage: 77.01% (+0.01%)

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: 24c88f2 | Docs | Datadog PR Page | Give us feedback!}

[thomas-lebeau]

Can you edit the PR title to follow our convention. This help generate clean changelog and release notes.

other than that, LGTM 👍