When priority queuing is implemented, lower-priority requests are typically kept on hold while higher-priority requests are served. The lower-priority requests may therefore be delayed if there is a constant flow of higher-priority requests.
To prevent delays for low-priority requests across multiple priority levels, you can configure weighted queuing for serving requests. The default weights for the priorities are:
You assign the minimum weight, zero (0), to requests that the NetScaler appliance should send to the server only if no requests are stored in any of the other queues. You assign the maximum weight, 101, to requests that the appliance should send to the server immediately, ahead of any requests stored in any of the other queues. Weights between these two set the relative priority of a particular queue in relation to the other queues. Queues with a higher weight are processed first; queues with a lower weight after the others have been processed. To assign the weights, see "Configuring a Priority Queuing Policy."