What would be cool if this somehow could do a comparison by provider. E.g. in the last outages anthropic models running on vertex were apparently less affected than those deployed elsewhere. (Not saying that one is better than the other, but would be a neat read out).