Hi! Could you support the Bencher metric format to enable end users to continously benchmark things with rlt? This would make rlt a great fit for CI driven benchmarks.

https://bencher.dev/docs/how-to/track-custom-benchmarks/