We should start tracking the run time performance of axe-core over time, that way we know ahead of time when [the performance degrades](https://github.com/dequelabs/axe-core/issues/5298]. The simplest way to do that would be to have a script run axe-core against a benchmark page a few times, take the average time, then append the result and date to a file in axe-core.
In terms of running it, we wouldn't want to run it for every pr, but we would want to track performance regularly and be able to tie changes to performance to commits so we know what changed and when.
We should start tracking the run time performance of axe-core over time, that way we know ahead of time when [the performance degrades](https://github.com/dequelabs/axe-core/issues/5298]. The simplest way to do that would be to have a script run axe-core against a benchmark page a few times, take the average time, then append the result and date to a file in axe-core.
In terms of running it, we wouldn't want to run it for every pr, but we would want to track performance regularly and be able to tie changes to performance to commits so we know what changed and when.