The queue time metric measures how long a request spends waiting in the router’s queue before being picked up by a dyno for processing. It reflects how quickly your application can start handling incoming requests.
We provide two views of queue time:
Average: Shows the typical waiting time experienced by requests during the selected period.
Max: Highlights the longest wait time observed, useful for spotting spikes or bottlenecks.
High or increasing queue times usually indicate that dynos are overloaded or insufficient in number. If autoscaling is enabled, you may see queue time correlate with dyno count adjustments, as the system adds dynos to reduce waiting.