Replica count undershoots or overshoots when concurrency mismatches batch size

For Services using adaptive batching or continuous batching, setting concurrency to a value different from the batch size results in suboptimal autoscaling, causing either underutilization or request queueing

For Services with adaptive batching or continuous batching enabled, set the concurrency parameter to match the configured batch size. This aligns processing capacity with batch requirements and optimizes throughput. Review your batch configuration and update concurrency accordingly.