Only use reproducible exclude newer errors in the test suite

[zanieb]

No description provided.

[codspeed-hq]

Merging this PR will degrade performance by 49.53%

❌ 4 regressed benchmarks
✅ 1 untouched benchmark

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	WallTime	resolve_warm_airflow	516.1 ms	1,022.7 ms	-49.53%
❌	WallTime	resolve_warm_jupyter	44.8 ms	62.5 ms	-28.39%
❌	WallTime	resolve_warm_jupyter_universal	143.1 ms	175.8 ms	-18.58%
❌	Simulation	resolve_warm_jupyter_universal	211.8 ms	241 ms	-12.08%

---

_{Comparing zb/exclude-newer-test (f7bf6ba) with main (9c0eec6)}