Notice: Over the next few months, we're reorganizing the App Engine documentation site to make it easier to find content and better align with the rest of Google Cloud products. The same content will be available, but the navigation will now match the rest of the Cloud products. If you have feedback or questions as you navigate the site, click Send Feedback.

Python 2 is no longer supported by the community. We recommend that you migrate Python 2 apps to Python 3.

Creating Task Handlers

This page describes how to create a task handler, the code that handles a push task. You must provide a request handler to process the task. The mapping from the request URL to the appropriate handler is declared in your service's app.yaml, just like any other request handler. Because you control how to map task requests to a handler, you're free to organize your task handlers. If your application processes many different kinds of tasks, you can add all the handlers to a single service, or you can distribute them among multiple services.

Writing a push task request handler

In the queue, the Task Queue service creates an HTTP header and sends it to an instance of the worker service specified by the task's target. Task Queue requests are sent from the IP address

Your handler does not need to be written in the same language that created and enqueued the task if the handler is in a separate service.

When you write your handler, follow these guidelines:

  • The code must return an HTTP status code within the range 200–299 to indicate success. Any other code indicates that the task failed.

  • Push tasks have a fixed completion deadline that depends on the scaling type of the service that's running them. Automatic scaling services must finish before 10 minutes have elapsed. Manual and basic scaling services can run up to 24 hours. If your handler misses the deadline, the Task Queue service assumes the task failed and will retry it.

    When a task's execution time nears the deadline, App Engine raises a DeadlineExceededError (from the module google.appengine.runtime) before the deadline is reached, so you can save your work or log whatever progress was made.

  • The handler must be idempotent. App Engine's Task Queue API is designed to provide "at least once" delivery; that is, if a task is successfully added, App Engine will deliver it to a handler at least once. Note that in some rare circumstances, multiple task execution is possible, so your code must ensure that there are no harmful side-effects of repeated execution.

The following example demonstrates retrieving an integer value from a request and adding the value to a counter that is maintained in Cloud Datastore:

from google.appengine.ext import ndb
import webapp2

COUNTER_KEY = 'default counter'

class Counter(ndb.Model):
    count = ndb.IntegerProperty(indexed=False)

class UpdateCounterHandler(webapp2.RequestHandler):
    def post(self):
        amount = int(self.request.get('amount'))

        # This task should run at most once per second because of the datastore
        # transaction write throughput.
        def update_counter():
            counter = Counter.get_or_insert(COUNTER_KEY, count=0)
            counter.count += amount


app = webapp2.WSGIApplication([
    ('/update_counter', UpdateCounterHandler)
], debug=True)

The mapping from the task's url /update-counter to the class UpdateCounterHandler is done inside WSGIApplication.

The worker.yaml file creates a service named "worker" and adds the worker code to it. Note that the handler's URL is secure because it specifies login:admin:

runtime: python27
api_version: 1
threadsafe: true
service: worker

- url: /.*
  login: admin

Task Queue uses the HTTP code in the handler's response to determine if the task succeeded. The response from the handler is seen only by the Task Queue service and only to determine if the task succeeded. The queue ignores all other fields in the response. Then the service discards the response. The originating app never receives any of the data. If a task fails, the Task Queue service retries the task by sending another request.

User-supplied data can be delivered to the handler in the request as a query string or as a payload in the request body. Inserting user data is described in Creating Tasks. If the request includes data, the handler must know how it was inserted into the request. The exact code you use to fetch the data from the request depends on the particular web framework you're using.

To test a task handler, sign in as an administrator and visit the handler's URL in your browser.

Reading request headers

A push task HTTP request has special headers set by App Engine, which contain task-specific information your handler can use.

If these headers are present in an external user request to your app, they are stripped and replaced. The sole exception is for requests from logged-in administrators of the application, who are allowed to set headers for testing purposes. On the other hand, headers are not removed when your app is running in the development server.

Requests from Task Queue will always contain the following headers:

Header Description
X-Appengine-QueueName The name of the queue (possibly "default" for the default push queue).
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 seconds since January 1st 1970.

If your request handler finds any of the headers listed above, it can trust that the request is a Task Queue request.

In addition, requests from Task Queue can contain the following headers:

Header Description
X-Appengine-TaskPreviousResponse The HTTP response code from the previous retry.
X-Appengine-TaskRetryReason The reason for retrying the task.
X-Appengine-FailFast Indicates that a task running fails immediately if an existing instance is not available.

Securing task handler URLs

If a task performs sensitive operations (such as modifying data), you might want to secure the handler URL to prevent a malicious external user from calling it directly. You can prevent users from accessing task URLs by restricting access to App Engine administrators. Task requests themselves are issued by App Engine and can always target a restricted URL.

You can restrict a URL by adding the login: admin element to the handler configuration in your app.yaml file.

For example:

- url: /your-task
  login: admin

What's next