This package allows you to use Google Cloud Tasks as your queue driver.
Using Cloud Tasks as a Laravel queue driver is fundamentally different than other Laravel queue drivers, like Redis.
Typically a Laravel queue has a worker that listens to incoming jobs using the queue:work / queue:listen command.
With Cloud Tasks, this is not the case. Instead, Cloud Tasks will schedule the job for you and make an HTTP request to your application with the job payload. There is no need to run a queue:work/listen command.
For more information on how to configure the Cloud Tasks queue, read the next section Configuring the queue
This package uses the HTTP request handler and doesn't support AppEngine. But feel free to contribute!
This package requires Laravel 5.6 or higher.
Please check the table below for supported Laravel and PHP versions:
| Laravel Version | PHP Version |
|---|---|
| 5.6 | 7.2 or 7.3 |
| 5.7 | 7.2 or 7.3 |
| 5.8 | 7.2 or 7.3 or 7.4 |
| 6.x | 7.2 or 7.3 or 7.4 or 8.0 |
| 7.x | 7.2 or 7.3 or 7.4 or 8.0 |
| 8.x | 7.3 or 7.4 or 8.0 |
(1) Require the package using Composer
composer require stackkit/laravel-google-cloud-tasks-queueOfficial documentation - Creating Cloud Tasks queues
(2) Add a new queue connection to config/queue.php
'cloudtasks' => [
'driver' => 'cloudtasks',
'project' => env('STACKKIT_CLOUD_TASKS_PROJECT', ''),
'location' => env('STACKKIT_CLOUD_TASKS_LOCATION', ''),
'handler' => env('STACKKIT_CLOUD_TASKS_HANDLER', ''),
'queue' => env('STACKKIT_CLOUD_TASKS_QUEUE', 'default'),
'service_account_email' => env('STACKKIT_CLOUD_TASKS_SERVICE_EMAIL', ''),
],
(3) Update the QUEUE_CONNECTION environment variable
QUEUE_CONNECTION=cloudtasks
(4) [Laravel ^8.0 and above only] configure failed tasks to use the database-uuids driver in config/queue.php
'failed' => [
'database' => env('DB_CONNECTION', 'mysql'),
'table' => 'failed_jobs',
'driver' => 'database-uuids',
],
(5) Create a new Cloud Tasks queue using gcloud
gcloud tasks queues create [QUEUE_ID]Now that the package is installed, the final step is to set the correct environment variables.
Please check the table below on what the values mean and what their value should be.
| Environment variable | Description | Example |
|---|---|---|
STACKKIT_CLOUD_TASKS_PROJECT |
The project your queue belongs to. | my-project |
STACKKIT_CLOUD_TASKS_LOCATION |
The region where the AppEngine is hosted | europe-west6 |
STACKKIT_CLOUD_TASKS_HANDLER |
The URL that Cloud Tasks will call to process a job. This should be the URL to your Laravel app with the handle-task path added |
https://<your website>.com/handle-task |
STACKKIT_CLOUD_TASKS_QUEUE |
The queue a job will be added to | emails |
STACKKIT_CLOUD_TASKS_SERVICE_EMAIL |
The email address of the AppEngine service account. Important, it should have the Cloud Tasks Enqueuer role. This is used for securing the handler. | my-service-account@appspot.gserviceaccount.com |
Set the GOOGLE_APPLICATION_CREDENTIALS environment variable with a path to the credentials file.
More info: https://cloud.google.com/docs/authentication/production
When you first create a queue using gcloud tasks queues create, the default settings will look something like this:
rateLimits:
maxBurstSize: 100
maxConcurrentDispatches: 1000
maxDispatchesPerSecond: 500.0
retryConfig:
maxAttempts: 100
maxBackoff: 3600s
maxDoublings: 16
minBackoff: 0.100s
Max burst size limits how fast tasks in queue are processed when many tasks are in the queue and the rate is high.
The maximum number of concurrent tasks that Cloud Tasks allows to be dispatched for this queue
The maximum rate at which tasks are dispatched from this queue.
Number of attempts per task. Cloud Tasks will attempt the task max_attempts times (that is, if the first attempt fails, then there will be max_attempts - 1 retries). Must be >= -1.|
A task will be scheduled for retry between min_backoff and max_backoff duration after it fails
The time between retries will double max_doublings times.
A task's retry interval starts at min_backoff, then doubles max_doublings times, then increases linearly, and finally retries retries at intervals of max_backoff up to max_attempts times.
For example, if min_backoff is 10s, max_backoff is 300s, and max_doublings is 3, then the a task will first be retried in 10s. The retry interval will double three times, and then increase linearly by 2^3 * 10s. Finally, the task will retry at intervals of max_backoff until the task has been attempted max_attempts times. Thus, the requests will retry at 10s, 20s, 40s, 80s, 160s, 240s, 300s, 300s, ....
To simulate a single queue:work/queue:listen process, simply set the maxConcurrentDispatches to 1:
gcloud tasks queues update [QUEUE_ID] --max-concurrent-dispatches=1
More information on configuring queues:
https://cloud.google.com/tasks/docs/configuring-queues
The job handler requires each request to have an OpenID token. In the installation step we set the service account email, and with that service account, Cloud Tasks will generate an OpenID token and send it along with the job payload to the handler.
This package verifies that the token is digitally signed by Google. Only Google Tasks will be able to call your handler.
More information about OpenID Connect:
https://developers.google.com/identity/protocols/oauth2/openid-connect