"Service latency metrics" docs

In Temporal Cloud metrics reference | Temporal Documentation, if I understand correctly, the latency metrics relate purely to API calls for those operations, rather than concepts like a workflow’s schedule-to-start duration.

Can this be clarified? Most readers are building their apps on top of the platform, and our thinking naturally goes to latencies as they relate to our workload.

latency metrics relate purely to API calls for those operations

you are correct

rather than concepts like a workflow’s schedule-to-start duration.

schedule to start latency as seen from service side is not exposed via cloud metrics. you should use your worker (sdk) metrics:

workflow_task_schedule_to_start_latency
activity_schedule_to_start_latency