Problem
The Lighthouse CI action config in `.github/workflows/lighthouse.yml` uses `runs: 1`, producing a single noisy/flaky performance sample instead of a median-of-N.
Location
`.github/workflows/lighthouse.yml`
Impact
Single-run Lighthouse scores are known to vary significantly run-to-run due to system load/network variance, making the resulting numbers unreliable for tracking real regressions.
Suggested fix
Increase `runs` to at least 3 and use the median/average score for budget assertions.
Problem
The Lighthouse CI action config in `.github/workflows/lighthouse.yml` uses `runs: 1`, producing a single noisy/flaky performance sample instead of a median-of-N.
Location
`.github/workflows/lighthouse.yml`
Impact
Single-run Lighthouse scores are known to vary significantly run-to-run due to system load/network variance, making the resulting numbers unreliable for tracking real regressions.
Suggested fix
Increase `runs` to at least 3 and use the median/average score for budget assertions.