test(replay): Fix flaky flush test

[Lms24]

This PR fixes the flakiness of our replay event flushing test by slightly changing the test requirements/expectations:

• Disabled compression to avoid potential timing problems
• Instead of asserting that a flush happens within a certain time frame from the last flush, we ease up this expectation so that we only expect that another flush happens after the previous one, given that click events happened in between.
• Instead of forcing a flush by triggering a visibility change (like we do a lot in other tests to avoid flakes), this test focuses on the max delay of flushing, thereby ensuring that our flushes are not debounced to eternity.

Note: I tested that flakes were gone by running this test 75x per CI job for multiple times and didn't see this test failing anymore.

closes#7266

[github-actions]

size-limit report 📦

Path	Size
@sentry/browser - ES5 CDN Bundle (gzipped + minified)	20.11 KB (0%)
@sentry/browser - ES5 CDN Bundle (minified)	62.49 KB (0%)
@sentry/browser - ES6 CDN Bundle (gzipped + minified)	18.74 KB (0%)
@sentry/browser - ES6 CDN Bundle (minified)	55.5 KB (0%)
@sentry/browser - Webpack (gzipped + minified)	20.48 KB (-0.01% 🔽)
@sentry/browser - Webpack (minified)	66.94 KB (0%)
@sentry/react - Webpack (gzipped + minified)	20.51 KB (0%)
@sentry/nextjs Client - Webpack (gzipped + minified)	48.04 KB (-0.01% 🔽)
@sentry/browser + @sentry/tracing - ES5 CDN Bundle (gzipped + minified)	27.04 KB (0%)
@sentry/browser + @sentry/tracing - ES6 CDN Bundle (gzipped + minified)	25.29 KB (-0.01% 🔽)
@sentry/replay ES6 CDN Bundle (gzipped + minified)	42.86 KB (+0.2% 🔺)
@sentry/replay - Webpack (gzipped + minified)	36.94 KB (+0.32% 🔺)
@sentry/browser + @sentry/tracing + @sentry/replay - ES6 CDN Bundle (gzipped + minified)	60.45 KB (+0.18% 🔺)
@sentry/browser + @sentry/replay - ES6 CDN Bundle (gzipped + minified)	53.99 KB (+0.17% 🔺)

[Lms24]

I'm actually not sure if there isn't some sort of problem with our flushing implementation rather than with test flakiness. We set maxFlushDelay to 500ms in init.js but very frequently, times between flushes seem to be closer to 1000ms. Sure, I guess there could be some wiggle room because of the event loop when the max timeout of the debounced flush resolves but almost double the actual time seems a little strange to me.

[Lms24]

Hmm we're even getting outliers with delays >2s: https://github.com/getsentry/sentry-javascript/actions/runs/4254985922/jobs/7402098735#step:11:264

It might just be that the runners are so slow that this just keeps on happening despite the maxdelay we set 🤔