Using Push Queues in Java
In App Engine push queues, a task is a unit of work to be performed by the application. Each task is an object of the TaskOptions class. Each Task object contains an application-specific URL with a request handler for the task, and an optional data payload that parameterizes the task.
For example, consider a calendaring application that needs to notify an invitee, via email, that an event has been updated. The data payload for this task consists of the email address and name of the invitee, along with a description of the event. The webhook might live at
/app_worker/send_email and contain a function that adds the relevant strings to an email template and sends the email. The app can create a separate task for each email it needs to send.
You can use push queues only within the App Engine environment; if you need to access App Engine tasks from outside of App Engine, use pull queues.
- Using push queues
- Push task execution
- Deferred tasks
- URL endpoints
- Push queues and the development server
- Push queues and backends
- Quotas and limits for push queues
Using push queues
A Java app sets up queues using a configuration file named
queue.xml, in the
WEB-INF/ directory inside the WAR. See Java Task Queue Configuration. Every app has a push queue named
default with some default settings.
To enqueue a task, you get a
Queue using the
QueueFactory, then call its
add() method. You can get a named queue specified in the
queue.xml file using the
getQueue() method of the factory, or you can get the default queue using
getDefaultQueue(). You can call the
add() method with a
TaskOptions instance (produced by
TaskOptions.Builder), or you can call it with no arguments to create a task with the default options for the queue.
The following code adds a task to a queue with options:
import com.google.appengine.api.taskqueue.Queue; import com.google.appengine.api.taskqueue.QueueFactory; import static com.google.appengine.api.taskqueue.TaskOptions.Builder.*; // ... Queue queue = QueueFactory.getDefaultQueue(); queue.add(withUrl("/worker").param("key", key));
Tasks added to this queue will execute by calling the request handler at the URL
/worker with the parameter
key. They will execute at the rate set in the
queue.xml file, or the default rate of 5 tasks per second.
Push task execution
App Engine executes push tasks by sending HTTP requests to your app. Specifying a programmatic asynchronous callback as an HTTP request is sometimes called a web hook. The web hook model enables efficient parallel processing.
The task's URL determines the handler for the task and the module that runs the handler.The handler is determined by the path part of the URL (the forward-slash separated string following the hostname), which is specified by the
urlparameter in the
TaskOptionsthat you include in your call to the
Queue.add()method. The url must be relative and local to your application's root directory. The module (or frontend or backend) and version in which the handler runs is determined by:
- The "Host"
headerparameter in the
TaskOptionsthat you include in your call to the
targetdirective in the
If you do not specify any of these parameters, the task will run in the same module/version in which it was enqueued, subject to these rules:
- If the default version of the app enqueues a task, the task will run on the default version. Note that if the app enqueues a task and the default version is changed before the task actually runs, the task will be executed in the new default version.
- If a non-default version enqueues a task, the task will always run on that same version.
The namespace in which a push task runs is determined when the task is added to the queue. By default, a task will run in the current namespace of the process that created the task. You can override this behavior by explicitly setting the namespace before adding a task to a queue, as described on the multitenancy page.
A task must finish executing and send an HTTP response value between 200–299 within 10 minutes of the original request. This deadline is separate from user requests, which have a 60-second deadline. If your task's execution nears the limit, App Engine raises a
DeadlineExceededException that you can catch to save your work or log progress before the deadline passes. If the task failed to execute, App Engine retries it based on criteria that you can configure.
Task request headers
Requests from the Task Queue service contain the following HTTP headers:
X-AppEngine-QueueName, the name of the queue (possibly
X-AppEngine-TaskName, the name of the task, or a system-generated unique ID if no name was specified
X-AppEngine-TaskRetryCount, the number of times this task has been retried; for the first attempt, this value is 0. This number includes attempts where the task failed due to a lack of available instances and never reached the execution phase.
X-AppEngine-TaskExecutionCount, the number of times this task has previously failed during the execution phase. This number does not include failures due to a lack of available instances.
X-AppEngine-TaskETA, the target execution time of the task, specified in milliseconds since January 1st 1970.
These headers are set internally by Google App Engine. If your request handler finds any of these headers, it can trust that the request is a Task Queue request. If any of the above headers are present in an external user request to your app, they are stripped. The exception being requests from logged in administrators of the application, who are allowed to set the headers for testing purposes.
Tasks may be created with the
X-AppEngine-FailFast header, which specifies that a task running on a backend fails immediately instead of waiting in a pending queue.
Google App Engine issues Task Queue requests from the IP address
The rate of task execution
You set the maximum processing rate for the entire queue when you configure the queue. App Engine uses a token bucket algorithm to execute tasks once they've been delivered to the queue. Each queue has a token bucket, and each bucket holds a certain number of tokens. Your app consumes a token each time it executes a task. If the bucket runs out of tokens, the system pauses until the bucket has more tokens. The rate at which the bucket is refilled is the limiting factor that determines the rate of the queue. See Defining Push Queues and Processing Rates for more details.
To ensure that the Task Queue system does not overwhelm your application, it may throttle the rate at which requests are sent. This throttled rate is known as the enforced rate. The enforced rate may be decreased when your application returns a 503 HTTP response code, or if there are no instances able to execute a request for an extended period of time. You can view the enforced rate on the Task Queue tab of the Administration Console.
The order of task execution
The order in which tasks are executed depends on several factors:
- The position of the task in the queue. App Engine attempts to process tasks based on FIFO (first in, first out) order. In general, tasks are inserted into the end of a queue, and executed from the head of the queue.
- The backlog of tasks in the queue. The system attempts to deliver the lowest latency possible for any given task via specially optimized notifications to the scheduler. Thus, in the case that a queue has a large backlog of tasks, the system's scheduling may "jump" new tasks to the head of the queue.
- The value of the task's etaMillis property. This specifies the earliest time that a task can execute. App Engine always waits until after the specified ETA to process push tasks.
- The value of the task's countdownMillis property. This specifies the minimum number of seconds to wait before executing a task. Countdown and eta are mutually exclusive; if you specify one, do not specify the other.
If a push task request handler returns an HTTP status code within the range 200–299, App Engine considers the task to have completed successfully. If the task returns a status code outside of this range, App Engine retries the task until it succeeds. The system backs off gradually to avoid flooding your application with too many requests, but schedules retry attempts for failed tasks to recur at a maximum of once per hour.
You can also configure your own scheme for task retries using the retry-parameters element in
When implementing the code for tasks (as worker URLs within your app), it is important to consider whether the task is idempotent. App Engine's Task Queue API is designed to only invoke a given task once; however, it is possible in exceptional circumstances that a task may execute multiple times (such as in the unlikely case of major system failure). Thus, your code must ensure that there are no harmful side-effects of repeated execution.
Setting up a handler for each distinct task (as described in the previous sections) can be cumbersome, as can serializing and deserializing complex arguments for the task — particularly if you have many diverse but small tasks that you want to run on the queue. The Java SDK includes an interface called DeferredTask. This interface lets you define a task as a single method. This interface uses Java serialization to package a unit of work into a Task Queue. A simple return from that method is considered success. Throwing any exception from that method is considered a failure.
Push tasks reference their implementation via URL. For example, a task which fetches and parses an RSS feed might use a worker URL called
/app_worker/fetch_feed. You can specify this worker URL or use the default. In general, you can use any URL as the worker for a task, so long as it is within your application; all task worker URLs must be specified as relative URLs:
import com.google.appengine.api.taskqueue.Queue; import com.google.appengine.api.taskqueue.QueueFactory; import com.google.appengine.api.taskqueue.TaskOptions.Method; import com.google.appengine.api.taskqueue.TaskOptions.Builder.*; // ... Queue queue = QueueFactory.getDefaultQueue(); queue.add(withUrl("/path/to/my/worker")); queue.add(withUrl("/path?a=b&c=d").method(Method.GET));
If you do not specify a worker URL, the task uses a default worker URL named after the queue:
A queue's default URL is used if, and only if, a task does not have a worker URL of its own. If a task does have its own worker URL, then it is only invoked at the worker URL, never another. Once inserted into a queue, its url endpoint cannot be changed.
Securing URLs for tasks
If a task performs sensitive operations (such as modifying important data), you might want to secure its worker URL to prevent a malicious external user from calling it directly. You can prevent users from accessing URLs of tasks by restricting access to administrator accounts. Task queues can access admin-only URLs. You can read about restricting URLs at Security and Authentication. An example you would use in
web.xml to restrict everything starting with
/tasks/ to admin-only is:
<security-constraint> <web-resource-collection> <web-resource-name>tasks</web-resource-name> <url-pattern>/tasks/*</url-pattern> </web-resource-collection> <auth-constraint> <role-name>admin</role-name> </auth-constraint> </security-constraint>
For more on the format of
web.xml, see the documentation on the deployment descriptor.
To test a task web hook, sign in as an administrator and visit the URL of the handler in your browser.
Push queues and the development server
When your app is running in the development server, tasks are automatically executed at the appropriate time just as in production.
To disable automatic execution of tasks, set the following jvm flag:
You can examine and manipulate tasks from the developer console at:
To execute tasks, select the queue by clicking on its name, select the tasks to execute, and click Run Now. To clear a queue without executing any tasks, click Purge Queue.
The development server and the production server behave differently:
- The development server doesn't respect the
<bucket-size>attributes of your queues. As a result, tasks are executed as close to their ETA as possible. Setting a rate of 0 doesn't prevent tasks from being executed automatically.
- The development server doesn't retry tasks.
- The development server doesn't preserve queue state across server restarts.
Push queues and backends
Push tasks typically must finish execution within 10 minutes. If you have push tasks that require more time or computing resources to process, you can use App Engine Backends to process these tasks outside of the normal limits of App Engine applications. The following code sample demonstrates how to create a push task addressed to an instance
1 of a backend named
import com.google.appengine.api.taskqueue.Queue; import com.google.appengine.api.taskqueue.QueueFactory; import static com.google.appengine.api.taskqueue.TaskOptions.Builder.*; import com.google.appengine.api.backends.*; // ... queue.add(withUrl("/path/to/my/worker").param("key", key).header("Host", BackendServiceFactory.getBackendService().getBackendAddress("backend1", 1));
Quotas and limits for push queues
Enqueuing a task in a push queue counts toward the following quotas:
- Task Queue Stored Task Count
- Task Queue Stored Task Bytes
- Task Queue API Calls
The Task Queue Stored Task Bytes quota is configurable in queue.xml by setting
<total-storage-limit>. This quota counts towards your Stored Data (billable) quota.
Execution of a task counts toward the following quotas:
- Incoming Bandwidth
- Outgoing Bandwidth
The act of executing a task consumes bandwidth-related quotas for the request and response data, just as if the request handler were called by a remote client. When the task queue processes a task, the response data is discarded.
Once a task has been executed or deleted, the storage used by that task is reclaimed. The reclaiming of storage quota for tasks happens at regular intervals, and this may not be reflected in the storage quota immediately after the task is deleted.
The following limits apply to the use of push queues:
|Push Queue Limits|
|Maximum task size||100KB|
|Maximum number of active queues (not including the default queue)||Free apps: 10 queues, Billed apps: 100 queues|
|Queue execution rate||500 task invocations per second per queue|
|Maximum countdown/ETA for a task||30 days from the current date and time|
|Maximum number of tasks that can be added in a batch||100 tasks|
|Maximum number of tasks that can be added in a transaction||5 tasks|