Eight queues per port

Eight queues per port enables fine-grained traffic prioritization across egress interfaces, allowing you to differentiate up to eight classes of service per physical port. You can use this feature to reduce congestion-related packet loss and improve quality of service for latency-sensitive traffic by assigning distinct scheduling, shaping, and weighted random early detection (WRED) policies to each queue. By default, only four queues are enabled per port; you can explicitly enable eight queues at the chassis level using the max-queues-per-interface option. The goal is to give you greater control over how traffic is classified, scheduled, and shaped, ensuring that high-priority traffic receives the appropriate treatment.