Inference Endpoints analytics real-time metrics, auto-refresh, and replica lifecycle view

AI Impact Summary

The Inference Endpoints analytics feature has been refreshed with real-time metrics, faster data loading, and a new replica lifecycle view. This enables operators to monitor request latency, error rates, and per-replica state transitions as they occur, improving troubleshooting for high-traffic endpoints. With customizable time ranges and auto-refresh, teams can track short-term spikes and long-term trends, reducing MTTR and enabling proactive capacity planning for ML inference workloads.