<queue-entries>
<!--Set the number of max concurrent requests to 10-->  
<queue>    
     <name>optimize-queue</name>                 <rate>20/s</rate>  
     <bucket-size>40</bucket-size>      
   <max-concurrent-requests>10</max-concurrent-requests>     </queue> </queue-entries> 

 Sample Code
This is a very straightforward example. As I said before, task queues are basically a URL handler. In this servlet, the GET will handle enqueueing a task. The task will POST to this same servlet and execute the doPost( ) method carrying out the task. In this case, it's just a simple counter. Notice the counter is a volatile property. If you access this servlet as GET request, it will enqueue another task. So, you will see the counter being incremented by both tasks.

public class TaskQInfo extends HttpServlet {

   private static volatile int TASK_COUNTER = 0;


   // Executed by user menu click
   public void doGet(HttpServletRequest req, HttpServletResponse resp)
        throws IOException {

       // Build a task using the TaskOptions Builder pattern from ** above
       Queue queue = QueueFactory.getDefaultQueue();
       queue.add(withUrl("/taskq_demo").method(TaskOptions.Method.POST)); 

       resp.getWriter().println( 
             "Task have been added to default queue...
");
       
       resp.getWriter().println(
             "Refresh this page to add another count task");
}

   // Executed by TaskQueue
   @Override
    protected void doPost(HttpServletRequest req, HttpServletResponse resp)
       throws ServletException, IOException {
       
       // This is the body of the task
       for(int i = 0; i < 1000; i++) {
             log.info("Processing: " + req.getHeader("X-AppEngine-TaskName") + "-" +           
               TASK_COUNTER++); 
              try { 
                 // Sleep for a second (if the rate is set to 1/s this will allow at 
                           most 1 more task to be processed)
                 Thread.sleep(1000); 
              } catch (InterruptedException e) { // ignore}
       }
    }
}

Task queues allow you to achieve some level of concurrency in your application by invoking background processes on demand. For very lengthy tasks, you might want to take a look at App Engine backends, which are basically special App Engine instances with no request time limit.