[madhu_ghalame]

Consider publishing latency, throughput, and cost metrics under different workloads to help teams make informed decisions.

[fionaattelnyx]

We'll be pushing benchmarks over the next few days, will update here when we have them