modelLatency

The expected model latency at maximum invocation per minute for the instance.