Agree on and document what should be performance tested

People doing development on FDB will need not only correctness testing, but performance testing. Performance validation is done internally within Apple’s and Snowflake’s development teams separately and differently, and a common agreement should be reached about what metrics should be evaluated under what conditions, and what is or is not considered a performance regression. The results of this would ideally be in some form externally publishable to give the community an idea of what is actually verified for each release, and so that any other developers will know what performance requirements their code will be required to meet.