Requests completed per second.
Percentage of requests that resulted in an error (0 to 1).
P95 latency in milliseconds.
P99 latency in milliseconds.
Number of rejections due to a full queue during the period.
Number of rejections due to queue timeouts during the period.
Process CPU usage (0 to 1).
Process RSS memory usage (0 to 1).
HSM Health (connectivity and responsiveness).
Health of Native Workers (Dynamonh cases).
Observability pipeline health (metrics, logs, traces).
A snapshot of the pool's metrics at a specific point in time, used as input for the autoscaling decision.