Examples provided herein describe a method for facilitating detection of microbursts in queues. For example, a physical processor of a computing device may dynamically determine, for each queue of a plurality of queues of a network switch, whether a monitoring threshold based on an amount of usage of a buffer memory by the plurality of queues. The physical processor may detect, for each queue, whether congestion exists based on whether throughput on the queue exceeds the determined monitoring threshold. The first physical processor may then report information about a set of queues experiencing microbursts in the network switch based on the detection of congestion for each queue.