Migrating push queues to Cloud Tasks (Java)

This page describes how you can migrate push queue code from Task Queues to Cloud Tasks. Cloud Tasks is now the preferred way of working with App Engine push queues.

With Cloud Tasks, you access the same service that you access with the Task Queues RPC API. This means that you don't need to recreate your existing push queues and push tasks. However, you must migrate code that creates or interacts with push queues or push tasks in order to use the Cloud Tasks API.

You can create and interact with push queues and push tasks using the Cloud Tasks REST and RPC APIs, the Cloud Tasks client library, the Google Cloud CLI, and the Google Cloud console. This page provides examples using the gcloud CLI and the Cloud Tasks client library.

In Cloud Tasks, all queues operate as push queues. In the remainder of this guide and in the Cloud Tasks documentation, the term queue is equivalent to the term push queue. Similarly, the term task is equivalent to the term push task.

Features not available in Cloud Tasks

The following features are not available in Cloud Tasks:

• Enqueueing tasks in Datastore transactions
• Using the deferred tasks library instead of a worker service
• Working with tasks in multi-tenant applications
• Simulating with the local development server
• Adding tasks asynchronously

Pricing and quotas

Migrating your push queues to Cloud Tasks might affect the pricing and quotas for your app.

Pricing

Cloud Tasks has its own pricing. As with Task Queues, sending requests to your App Engine app with a task can cause your app to incur costs.

Quotas

The Cloud Tasks quotas are different from the quotas for Task Queues. Like with Task Queues, sending requests to your App Engine app from Cloud Tasks might impact your App Engine request quotas.

Before migrating

The following sections discuss the setup steps before migrating your push queues to Cloud Tasks.

Migrating pull queues

Before you begin, migrate pull queues before following the instructions in this guide to migrate push queues. Migrating pull queues after migrating push queues is not recommended because the required use of the queue.yaml file is likely to cause unexpected behavior with Cloud Tasks.

Protecting queue configuration

Once you begin the process of migrating to Cloud Tasks, modifying your queue.yaml file can cause unexpected behavior and is not recommended. Protect your queue configuration from modifications by the queue.yaml file with the following steps.

1. Configure the gcloud CLI to omit your queue.yaml file in future deployments.

   Add your queue.yaml file to a .gcloudignore file. To check if you already have a .gcloudignore file, you can run the following command in your terminal from the top level directory of your app. This command will output the filename if the file exists.

   ls -a | grep .gcloudignore

   To learn more about .gcloudignore files, read the .gcloudignore reference.

2. Restrict permissions on your queue.yaml file.

   Follow the best practices described in our guide on securing queue configuration.

3. Learn about Cloud Tasks and the queue.yaml file (optional).

   When using the Cloud Tasks API to manage your queue configuration, deploying a queue.yaml file overrides the configuration set by Cloud Tasks, which can cause unexpected behavior. Read Using queue management versus queue.yaml to learn more.

Enabling the Cloud Tasks API

To enable the Cloud Tasks API, click Enable on the Cloud Tasks API in the API Library. If you see a Manage button instead of an Enable button, you have previously enabled the Cloud Tasks API for your project and don't need to do so again.

Authenticating your app to the Cloud Tasks API

You must authenticate your app to the Cloud Tasks API. This section discusses authentication for two different use cases.

To develop or test your app locally, we recommend using a service account. For instructions on setting up a service account and connecting it to your app, read Obtaining and providing service account credentials manually.

To deploy your app on App Engine, you don't need to provide any new authentication. The Application Default Credentials (ADC) infer authentication details for App Engine apps.

Downloading the gcloud CLI

Download and install the gcloud CLI to use the gcloud CLI with the Cloud Tasks API if you have not installed it previously. Run the following command from your terminal if you already have the gcloud CLI installed.

gcloud components update

Importing the Cloud Client Libraries

To use the Cloud Tasks client library with your App Engine app:

1. Specify the Cloud Tasks client library dependency in your pom.xml file:

   <dependency>
     <groupId>com.google.cloud</groupId>
     <artifactId>google-cloud-tasks</artifactId>
     <version>1.3.0</version>
   </dependency>
   

2. Import the Cloud Tasks client library dependencies in the files that are responsible for creating and enqueueing your tasks:

   import com.google.cloud.tasks.v2.AppEngineHttpRequest;
   import com.google.cloud.tasks.v2.CloudTasksClient;
   import com.google.cloud.tasks.v2.HttpMethod;
   import com.google.cloud.tasks.v2.QueueName;
   import com.google.cloud.tasks.v2.Task;
   

Creating and managing queues

This section describes how to create and manage queues using the Cloud Tasks API.

With Cloud Tasks, you don't use a queue.yaml file to create or manage queues. Instead, you use the Cloud Tasks API. Using both a queue.yaml file and the Cloud Tasks API is not recommended, but it might be an inevitable part of migrating from Task Queues to Cloud Tasks depending on your app. Read Using Queue management versus queue.yaml to learn about best practices.

Creating queues

Read this section if your app creates queues programmatically, or if you want to create additional queues from the command line.

In Cloud Tasks, queue names have the form projects/PROJECT_ID/locations/LOCATION_ID/queues/QUEUE_ID. The LOCATION_ID portion of the queue name corresponds to a Google Cloud region. The QUEUE_ID portion of the queue name is equivalent to the Task Queues queue name field. The queue name is generated during queue creation based on the project, region and QUEUE_ID you specify.

In general, the queue location (i.e., region) must be the same as the region of your app. The two exceptions to this rule are for apps using the europe-west region and apps using the us-central region. In Cloud Tasks, these regions are called europe-west1 and us-central1 respectively.

You can specify optional queue configuration during queue creation, however you can also do so by updating the queue after it has been created.

You don't need to recreate existing queues. Instead, migrate the code that interacts with your existing queues by reading the relevant parts of this guide.

Note: The App Engine Task Queue service automatically creates a default queue named default. Cloud Tasks does not automatically create this queue.

Reusing queue names

You must wait 7 days after deleting a queue to create a queue with the same queue ID in the same project and location (i.e., region).

The following example creates two queues using Cloud Tasks. The first queue has queue ID queue-blue, and is configured to send all tasks to the version v2 of the service task-module at a rate of 5/s. The second queue has queue ID queue-red and dispatches tasks at a rate of 1/s. Both are created on the project with project ID your-project-id in location us-central1. This is the Cloud Tasks equivalent of creating queues in Task Queues.

gcloud

The gcloud CLI infers the project and location from the gcloud CLI configuration.

gcloud tasks queues create queue-blue \
--max-dispatches-per-second=5 \
--routing-override=service:task-module,version:v2

gcloud tasks queues create queue-red \
--max-dispatches-per-second=1

client library

import com.google.cloud.tasks.v2.AppEngineRouting;
import com.google.cloud.tasks.v2.CloudTasksClient;
import com.google.cloud.tasks.v2.LocationName;
import com.google.cloud.tasks.v2.Queue;
import com.google.cloud.tasks.v2.QueueName;
import com.google.cloud.tasks.v2.RateLimits;

public class CreateQueue {
  public static void createQueue(
      String projectId, String locationId, String queueBlueName, String queueRedName)
      throws Exception {
    try (CloudTasksClient client = CloudTasksClient.create()) {
      // TODO(developer): Uncomment these lines and replace with your values.
      // String projectId = "your-project-id";
      // String locationId = "us-central1";
      // String queueBlueName = "queue-blue";
      // String queueRedName = "queue-red";

      LocationName parent = LocationName.of(projectId, locationId);

      Queue queueBlue =
          Queue.newBuilder()
              .setName(QueueName.of(projectId, locationId, queueBlueName).toString())
              .setRateLimits(RateLimits.newBuilder().setMaxDispatchesPerSecond(5.0))
              .setAppEngineRoutingOverride(
                  AppEngineRouting.newBuilder().setVersion("v2").setService("task-module"))
              .build();

      Queue queueRed =
          Queue.newBuilder()
              .setName(QueueName.of(projectId, locationId, queueRedName).toString())
              .setRateLimits(RateLimits.newBuilder().setMaxDispatchesPerSecond(1.0))
              .build();

      Queue[] queues = new Queue[] {queueBlue, queueRed};
      for (Queue queue : queues) {
        Queue response = client.createQueue(parent, queue);
        System.out.println(response);
      }
    }
  }
}

To learn more, read the Cloud Tasks reference Creating a Cloud Tasks queue.

Setting the queue processing rate

The following table lists the fields that differ from Task Queues to Cloud Tasks.

Task Queues field	Cloud Tasks field	Description
rate	max_dispatches_per_second	The maximum rate at which tasks are dispatched from the queue
max_concurrent_requests	max_concurrent_dispatches	The maximum number of concurrent tasks that can be dispatched from the queue
bucket_size	max_burst_size	Cloud Tasks calculates a get-only property max_burst_size that limits how fast tasks in the queue are processed based on the value of max_dispatches_per_second. This field allows the queue to have a high rate so that processing starts shortly after a task is enqueued, but still limits resource usage when many tasks are enqueued in a short period of time. For App Engine queues that were created or updated using a queue.xml/yaml file, max_burst_size is initially equal to bucket_size. However, if the queue is later passed to an update command using any Cloud Tasks interface, max_burst_size will be reset based on the value of max_dispatches_per_second, regardless of whether max_dispatches_per_second is updated.
total_storage_limit	Deprecated in Cloud Tasks	Cloud Tasks does not support setting a custom storage limit

You can set the queue processing rate when you create the queue or update it afterwards. The following example uses Cloud Tasks to set the processing rate on a queue named queue-blue that has already been created. If queue-blue was created or configured using a queue.yaml file, the following example resetsmax_burst_size based on the max_dispatches_per_second value of 20. This is the Cloud Tasks equivalent of setting the queue processing rate in Task Queues.

gcloud

gcloud tasks queues update queue-blue \
--max-dispatches-per-second=20 \
--max-concurrent-dispatches=10

client library

import com.google.cloud.tasks.v2.CloudTasksClient;
import com.google.cloud.tasks.v2.LocationName;
import com.google.cloud.tasks.v2.Queue;
import com.google.cloud.tasks.v2.QueueName;
import com.google.cloud.tasks.v2.RateLimits;
import com.google.cloud.tasks.v2.UpdateQueueRequest;

public class UpdateQueue {
  public static void updateQueue(String projectId, String locationId, String queueId)
      throws Exception {
    try (CloudTasksClient client = CloudTasksClient.create()) {
      // TODO(developer): Uncomment these lines and replace with your values.
      // String projectId = "your-project-id";
      // String locationId = "us-central1";
      // String queueId = "queue-blue";

      LocationName parent = LocationName.of(projectId, locationId);

      Queue queueBlue =
          Queue.newBuilder()
              .setName(QueueName.of(projectId, locationId, queueId).toString())
              .setRateLimits(
                  RateLimits.newBuilder()
                      .setMaxDispatchesPerSecond(20.0)
                      .setMaxConcurrentDispatches(10))
              .build();

      UpdateQueueRequest request = UpdateQueueRequest.newBuilder().setQueue(queueBlue).build();

      Queue response = client.updateQueue(request);
      System.out.println(response);
    }
  }
}

To learn more, see Define rate limits.

Disabling and resuming queues

Cloud Tasks uses the term pause in the same way that Task Queues uses the term disable. Pausing a queue stops the tasks in the queue from executing until the queue is resumed. However, you can continue to add tasks to a queue that is paused. Cloud Tasks uses the term resume in the same way as Task Queues does.

The following example pauses a queue with queue ID queueName . This is the Cloud Tasks equivalent of disabling queues in Task Queues.

gcloud

gcloud tasks queues pause 

queueName

client library

import com.google.cloud.tasks.v2.CloudTasksClient;
import com.google.cloud.tasks.v2.QueueName;

public class PauseQueue {
  public static void pauseQueue(String projectId, String locationId, String queueId)
      throws Exception {
    try (CloudTasksClient client = CloudTasksClient.create()) {
      // TODO(developer): Uncomment these lines and replace with your values.
      // String projectId = "your-project-id";
      // String locationId = "us-central1";
      // String queueId = "foo";

      // Construct the fully qualified queue name.
      String queueName = QueueName.of(projectId, locationId, queueId).toString();

      client.pauseQueue(queueName);
      System.out.println("Queue Paused.");
    }
  }
}

To learn more, read the Cloud Tasks reference Pausing queues.

Deleting queues

Once you delete a queue, you must wait 7 days before creating a queue with the same name. Consider purging all tasks from a queue and reconfiguring the queue if you cannot wait 7 days.

The following example deletes the queue with queue ID foo . This is the Cloud Tasks equivalent of deleting queues in Task Queues.

gcloud

gcloud tasks queues delete 

foo

client library

import com.google.cloud.tasks.v2.CloudTasksClient;
import com.google.cloud.tasks.v2.QueueName;

public class DeleteQueue {
  public static void deleteQueue(String projectId, String locationId, String queueId)
      throws Exception {
    try (CloudTasksClient client = CloudTasksClient.create()) {
      // TODO(developer): Uncomment these lines and replace with your values.
      // String projectId = "your-project-id";
      // String locationId = "us-central1";
      // String queueId = "foo";

      // Construct the fully qualified queue name.
      String queueName = QueueName.of(projectId, locationId, queueId).toString();

      client.deleteQueue(queueName);
      System.out.println("Queue Deleted.");
    }
  }
}

To learn more, read the Cloud Tasks reference Deleting queues.

Creating and managing tasks

This section describes how to create and manage tasks using the Cloud Tasks API.

Creating tasks

The following table lists the fields that differ from Task Queues to Cloud Tasks.

Task Queues field	Cloud Tasks field	Description
NEW in Cloud Tasks	app_engine_http_request	Creates a request that targets an App Engine service. These tasks are referred to as App Engine tasks.
method	http_method	Specifies the request method; for example, POST
url	relative_uri	Specifies the task handler. Note the difference in the final letter: i for uniform resource identifier rather than l for uniform resource locator
target	app_engine_routing	Optional. Specifies the App Engine service, version, and instance for an App Engine task. If not set, the default service, version, and instance are used.

The following example creates a task that routes to the handler /worker on the default App Engine service. This is the Cloud Tasks equivalent to creating tasks in Task Queues.

gcloud

gcloud tasks create-app-engine-task --queue=default \
--method=POST --relative-uri=/worker?key=key

client library

import com.google.cloud.tasks.v2.AppEngineHttpRequest;
import com.google.cloud.tasks.v2.CloudTasksClient;
import com.google.cloud.tasks.v2.HttpMethod;
import com.google.cloud.tasks.v2.QueueName;
import com.google.cloud.tasks.v2.Task;
import com.google.protobuf.ByteString;
import java.nio.charset.Charset;

public class CreateTask {
  public static void createTask(String projectId, String locationId, String queueId)
      throws Exception {
    try (CloudTasksClient client = CloudTasksClient.create()) {
      // TODO(developer): Uncomment these lines and replace with your values.
      // String projectId = "your-project-id";
      // String locationId = "us-central1";
      // String queueId = "default";
      String key = "key";

      // Construct the fully qualified queue name.
      String queueName = QueueName.of(projectId, locationId, queueId).toString();

      // Construct the task body.
      Task taskParam =
          Task.newBuilder()
              .setAppEngineHttpRequest(
                  AppEngineHttpRequest.newBuilder()
                      .setRelativeUri("/worker?key=" + key)
                      .setHttpMethod(HttpMethod.GET)
                      .build())
              .build();

      Task taskPayload =
          Task.newBuilder()
              .setAppEngineHttpRequest(
                  AppEngineHttpRequest.newBuilder()
                      .setBody(ByteString.copyFrom(key, Charset.defaultCharset()))
                      .setRelativeUri("/worker")
                      .setHttpMethod(HttpMethod.POST)
                      .build())
              .build();

      // Send create task request.
      Task[] tasks = new Task[] {taskParam, taskPayload};
      for (Task task : tasks) {
        Task response = client.createTask(queueName, task);
        System.out.println(response);
      }
    }
  }
}

To learn more, read the Cloud Tasks reference Creating App Engine tasks.

Specifying the target service and routing

Specifying the App Engine target service, version, and instance for App Engine tasks is optional. By default, App Engine tasks are routed to the service, version, and instance that are the default at the time the task is attempted.

Set the task's app_engine_routing property during task creation to specify a different App Engine service, version, or instance for your task.

To route all tasks on a given queue to the same App Engine service, version, and instance, you can set the app_engine_routing_override property on the queue.

To learn more, read the Cloud Tasks reference Configure routing.

Passing data to the handler

As with Task Queues, you can pass data to the handler in two ways using Cloud Tasks. You can either pass data as query parameters in the relative URI, or you can pass data in the request body using the HTTP methods POST or PUT.

Cloud Tasks uses the term body in the same way that Task Queues uses the term payload. In Cloud Tasks, the default body content type is octet-stream rather than plain text. You can set the body content type by specifying it in the header.

The following example passes a key to the handler /worker in two different ways. This is the Cloud Tasks equivalent of passing data to the handler in Task Queues.

console

gcloud tasks create-app-engine-task --queue=default --method=GET  \
--relative-uri=

/worker

?key=blue --routing=service:worker

gcloud tasks create-app-engine-task --queue=default --method=POST \
--relative-uri=

/worker

 --routing=service:worker \
--body-content="{'key': 'blue'}"

client library

import com.google.cloud.tasks.v2.AppEngineHttpRequest;
import com.google.cloud.tasks.v2.CloudTasksClient;
import com.google.cloud.tasks.v2.HttpMethod;
import com.google.cloud.tasks.v2.QueueName;
import com.google.cloud.tasks.v2.Task;
import com.google.protobuf.ByteString;
import java.nio.charset.Charset;

public class CreateTask {
  public static void createTask(String projectId, String locationId, String queueId)
      throws Exception {
    try (CloudTasksClient client = CloudTasksClient.create()) {
      // TODO(developer): Uncomment these lines and replace with your values.
      // String projectId = "your-project-id";
      // String locationId = "us-central1";
      // String queueId = "default";
      String key = "key";

      // Construct the fully qualified queue name.
      String queueName = QueueName.of(projectId, locationId, queueId).toString();

      // Construct the task body.
      Task taskParam =
          Task.newBuilder()
              .setAppEngineHttpRequest(
                  AppEngineHttpRequest.newBuilder()
                      .setRelativeUri("/worker?key=" + key)
                      .setHttpMethod(HttpMethod.GET)
                      .build())
              .build();

      Task taskPayload =
          Task.newBuilder()
              .setAppEngineHttpRequest(
                  AppEngineHttpRequest.newBuilder()
                      .setBody(ByteString.copyFrom(key, Charset.defaultCharset()))
                      .setRelativeUri("/worker")
                      .setHttpMethod(HttpMethod.POST)
                      .build())
              .build();

      // Send create task request.
      Task[] tasks = new Task[] {taskParam, taskPayload};
      for (Task task : tasks) {
        Task response = client.createTask(queueName, task);
        System.out.println(response);
      }
    }
  }
}

Naming tasks

Specifying the task name is optional. If you don't specify the task name, Cloud Tasks constructs it for you by generating a task ID and inferring the project and location (i.e., region) based on the queue you specified during task creation.

Task names have the form projects/PROJECT_ID/locations/LOCATION_ID/queues/QUEUE_ID/tasks/TASK_ID. The TASK_ID portion of the task name is equivalent to the Task Queues task name field.

Reusing task names

You must wait before reusing the name of a task. The amount of time you must wait before doing so differs based on whether the queue dispatching the task was created in Cloud Tasks or Task Queues.

For tasks on queues that were created using Task Queues (including the default queue), you must wait approximately 9 days after the original task was deleted or executed. For tasks on queues that were created using Cloud Tasks, you must wait approximately 1 hour after the original task was deleted or executed.

The following example creates a task with the TASK_ID set to first-try, and adds it to the default queue. This is the Cloud Tasks equivalent of naming tasks in Task Queues.

gcloud

The gcloud CLI constructs the task name by inferring the project and location from your configuration.

gcloud tasks create-app-engine-task first-try --queue=default \
--method=GET --relative-uri=

/worker

client library

With the client library, you must specify the full task name if you want to specify the TASK_ID. The project and location must match the project and location of the queue to which the task is added.

import com.google.cloud.tasks.v2.AppEngineHttpRequest;
import com.google.cloud.tasks.v2.CloudTasksClient;
import com.google.cloud.tasks.v2.HttpMethod;
import com.google.cloud.tasks.v2.QueueName;
import com.google.cloud.tasks.v2.Task;
import com.google.cloud.tasks.v2.TaskName;

public class CreateTaskWithName {
  public static void createTaskWithName(
      String projectId, String locationId, String queueId, String taskId) throws Exception {
    try (CloudTasksClient client = CloudTasksClient.create()) {
      // TODO(developer): Uncomment these lines and replace with your values.
      // String projectId = "your-project-id";
      // String locationId = "us-central1";
      // String queueId = "default";
      // String taskId = "first-try"

      String queueName = QueueName.of(projectId, locationId, queueId).toString();

      Task.Builder taskBuilder =
          Task.newBuilder()
              .setName(TaskName.of(projectId, locationId, queueId, taskId).toString())
              .setAppEngineHttpRequest(
                  AppEngineHttpRequest.newBuilder()
                      .setRelativeUri("/worker")
                      .setHttpMethod(HttpMethod.GET)
                      .build());

      // Send create task request.
      Task response = client.createTask(queueName, taskBuilder.build());
      System.out.println(response);
    }
  }
}

Retrying failed tasks

You can set task retry configuration on queues during queue creation, or by updating the queue. The following table lists the Task Queues field and the corresponding Cloud Tasks field.

Task Queues field	Cloud Tasks field
task_retry_limit	max_attempts
task_age_limit	max_retry_duration
min_backoff_seconds	min_backoff
max_backoff_seconds	max_backoff
max_doublings	max_doublings

Task-specific retry parameters

Task-specific retry parameters that were configured in Task Queues work in Cloud Tasks, however you cannot edit them or set them on new tasks. To change the retry parameters for a task that has task-specific retry parameters, recreate the task with a Cloud Tasks queue that has the desired retry parameters.

The following example demonstrates various retry scenarios:

• In fooqueue, tasks are retried up to seven times and for up to two days from the first execution attempt. After both limits are passed, it fails permanently.
• In barqueue, App Engine attempts to retry tasks, increasing the interval linearly between each retry until reaching the maximum backoff and retrying indefinitely at the maximum interval (so the intervals between requests are 10s, 20s, 30s, ..., 190s, 200s, 200s, ...).
• In bazqueue, the retry interval starts at 10s, then doubles three times, then increases linearly, and finally retries indefinitely at the maximum interval (so the intervals between requests are 10s, 20s, 40s, 80s, 160s, 240s, 300s, 300s, ...).

This is the Cloud Tasks equivalent of retrying tasks in Task Queues.

gcloud

When setting options that specify a number of seconds, you must include s after the integer number (e.g. 200s not 200).

gcloud tasks queues create fooqueue \
--max-attempts=7 \
--max-retry-duration=172800s  #2*60*60*24 seconds in 2 days

gcloud tasks queues create barqueue \
--min-backoff=10s \
--max-backoff=200s \
--max-doublings=0

gcloud tasks queues create bazqueue \
--min-backoff=10s \
--max-backoff=300s \
--max-doublings=3

client library

import com.google.cloud.tasks.v2.CloudTasksClient;
import com.google.cloud.tasks.v2.LocationName;
import com.google.cloud.tasks.v2.Queue;
import com.google.cloud.tasks.v2.QueueName;
import com.google.cloud.tasks.v2.RateLimits;
import com.google.cloud.tasks.v2.RetryConfig;
import com.google.protobuf.Duration;

public class RetryTask {
  public static void retryTask(
      String projectId, String locationId, String fooqueue, String barqueue, String bazqueue)
      throws Exception {
    try (CloudTasksClient client = CloudTasksClient.create()) {
      // TODO(developer): Uncomment these lines and replace with your values.
      // String projectId = "your-project-id";
      // String locationId = "us-central1";
      // String fooqueue = "fooqueue";
      // String barqueue = "barqueue";
      // String bazqueue = "bazqueue";

      LocationName parent = LocationName.of(projectId, locationId);

      Duration retryDuration = Duration.newBuilder().setSeconds(2 * 60 * 60 * 24).build();
      Duration min = Duration.newBuilder().setSeconds(10).build();
      Duration max1 = Duration.newBuilder().setSeconds(200).build();
      Duration max2 = Duration.newBuilder().setSeconds(300).build();

      Queue foo =
          Queue.newBuilder()
              .setName(QueueName.of(projectId, locationId, fooqueue).toString())
              .setRateLimits(RateLimits.newBuilder().setMaxDispatchesPerSecond(1.0))
              .setRetryConfig(
                  RetryConfig.newBuilder().setMaxAttempts(7).setMaxRetryDuration(retryDuration))
              .build();

      Queue bar =
          Queue.newBuilder()
              .setName(QueueName.of(projectId, locationId, barqueue).toString())
              .setRateLimits(RateLimits.newBuilder().setMaxDispatchesPerSecond(1.0))
              .setRetryConfig(
                  RetryConfig.newBuilder()
                      .setMinBackoff(min)
                      .setMaxBackoff(max1)
                      .setMaxDoublings(0))
              .build();

      Queue baz =
          Queue.newBuilder()
              .setName(QueueName.of(projectId, locationId, bazqueue).toString())
              .setRateLimits(RateLimits.newBuilder().setMaxDispatchesPerSecond(1.0))
              .setRetryConfig(
                  RetryConfig.newBuilder()
                      .setMinBackoff(min)
                      .setMaxBackoff(max2)
                      .setMaxDoublings(3))
              .build();

      Queue[] queues = new Queue[] {foo, bar, baz};
      for (Queue queue : queues) {
        Queue response = client.createQueue(parent, queue);
        System.out.println(response);
      }
    }
  }
}

To learn more, read the Cloud Tasks reference Set retry parameters.

Deleting tasks from a queue

When you delete a task, you must wait 9 days before creating a task with the same name if the task was on a queue created using a queue.yaml file, or 1 hour if the task was on a queue created using Cloud Tasks.

The following example deletes the task with task ID foo from the queue with queue ID queue1. This is the Cloud Tasks equivalent of deleting tasks in Task Queues.

gcloud

The task project and location are inferred from the gcloud CLI default project.

gcloud tasks delete foo --queue=queue1

client library

import com.google.cloud.tasks.v2.CloudTasksClient;
import com.google.cloud.tasks.v2.TaskName;

public class DeleteTask {
  public static void deleteTask(String projectId, String locationId, String queueId, String taskId)
      throws Exception {
    try (CloudTasksClient client = CloudTasksClient.create()) {
      // TODO(developer): Uncomment these lines and replace with your values.
      // String projectId = "your-project-id";
      // String locationId = "us-central1";
      // String queueId = "queue1";
      // String taskId = "foo";

      // Construct the fully qualified queue name.
      String taskName = TaskName.of(projectId, locationId, queueId, taskId).toString();

      client.deleteTask(taskName);
      System.out.println("Task Deleted.");
    }
  }
}

To learn more, read the Cloud Tasks reference Deleting a task from a queue.

Purging tasks

The following example purges all tasks from the queue with queue ID foo . This is the Cloud Tasks equivalent of purging tasks in Task Queues.

gcloud

The queue project and location are inferred from the gcloud CLI default project.

gcloud tasks queues purge 

foo

client library

import com.google.cloud.tasks.v2.CloudTasksClient;
import com.google.cloud.tasks.v2.QueueName;

public class PurgeQueue {
  public static void purgeQueue(String projectId, String locationId, String queueId)
      throws Exception {
    try (CloudTasksClient client = CloudTasksClient.create()) {
      // TODO(developer): Uncomment these lines and replace with your values.
      // String projectId = "your-project-id";
      // String locationId = "us-central1";
      // String queueId = "foo";

      // Construct the fully qualified queue name.
      String queueName = QueueName.of(projectId, locationId, queueId).toString();

      client.purgeQueue(queueName);
      System.out.println("Queue Purged.");
    }
  }
}

To learn more, read the Cloud Tasks reference Purging all tasks from a queue.

A Java 8 Cloud Tasks example

The following example is a basic setup for creating a queue and enqueueing a task to it with Cloud Tasks. It assumes that the developer has created a pom.xml file to specify the Cloud Tasks dependency as described in the section Importing the client library. This is the Cloud Tasks equivalent of a Java 8 task queue example in Task Queues.

The file that is responsible for creating and enqueueing the task creates a task and adds it to the default queue using the Cloud Tasks client library:

import com.google.cloud.tasks.v2.AppEngineHttpRequest;
import com.google.cloud.tasks.v2.CloudTasksClient;
import com.google.cloud.tasks.v2.HttpMethod;
import com.google.cloud.tasks.v2.QueueName;
import com.google.cloud.tasks.v2.Task;
import java.io.IOException;
import javax.servlet.ServletException;
import javax.servlet.annotation.WebServlet;
import javax.servlet.http.HttpServlet;
import javax.servlet.http.HttpServletRequest;
import javax.servlet.http.HttpServletResponse;

@WebServlet(
    name = "TaskEnqueue",
    description = "Enqueue a task targeted at endpoint '/cloudtasks/worker'",
    urlPatterns = "/cloudtasks/enqueue")
public class Enqueue extends HttpServlet {

  // TODO(developer): Replace these variables before running the sample.
  static final String projectId = "my-project-id";
  static final String locationId = "us-central1";

  // Function creates Cloud Tasks from form submissions.
  protected void doPost(HttpServletRequest request, HttpServletResponse response)
      throws ServletException, IOException {
    String key = request.getParameter("key");

    try (CloudTasksClient client = CloudTasksClient.create()) {
      // Construct the fully qualified queue name.
      String queueName = QueueName.of(projectId, locationId, "default").toString();

      // Construct the task body.
      Task task =
          Task.newBuilder()
              .setAppEngineHttpRequest(
                  AppEngineHttpRequest.newBuilder()
                      .setRelativeUri("/cloudtasks/worker?key=" + key)
                      .setHttpMethod(HttpMethod.POST)
                      .build())
              .build();

      // Add the task to the default queue.
      Task taskResponse = client.createTask(queueName, task);
      System.out.println("Task created: " + taskResponse.getName());
    }

    response.sendRedirect("/");
  }
}

The file that defines the worker handles the task:

import java.io.IOException;
import java.util.logging.Logger;
import javax.servlet.ServletException;
import javax.servlet.annotation.WebServlet;
import javax.servlet.http.HttpServlet;
import javax.servlet.http.HttpServletRequest;
import javax.servlet.http.HttpServletResponse;

@WebServlet(
    name = "TaskWorker",
    description = "Endpoint to process Cloud Task requests",
    urlPatterns = "/cloudtasks/worker"
)
public class Worker extends HttpServlet {

  private static final Logger log = Logger.getLogger(Worker.class.getName());

  // Worker function to process POST requests from Cloud Tasks targeted at the
  // '/cloudtasks/worker' endpoint.
  protected void doPost(HttpServletRequest request, HttpServletResponse response)
      throws ServletException, IOException {
    String key = request.getParameter("key");
    log.info("Worker is processing " + key);
  }
}

What's next

• Cloud Tasks documentation
• Cloud Tasks client library
• Cloud Tasks REST reference overview
• Cloud Tasks RPC reference overview