Replace celery[sqs] with celery[redis]

.github/workflows/checks.yml

@@ -9,7 +9,6 @@ env:
   DEBUG: True
   ANTIVIRUS_ENABLED: 0
   NOTIFY_ENVIRONMENT: test
-  NOTIFICATION_QUEUE_PREFIX: local_dev_10x
   STATSD_HOST: localhost
   SES_STUB_URL: None
   NOTIFY_APP_NAME: api

.github/workflows/daily_checks.yml

@@ -13,7 +13,6 @@ env:
   DEBUG: True
   ANTIVIRUS_ENABLED: 0
   NOTIFY_ENVIRONMENT: test
-  NOTIFICATION_QUEUE_PREFIX: local_dev_10x
   STATSD_HOST: localhost
   SES_STUB_URL: None
   NOTIFY_APP_NAME: api

README.md

@@ -49,7 +49,6 @@ NOTE: when you change .env in the future, you'll need to rebuild the devcontaine
 Things to change:
 
 - If you're not the first to deploy, only replace the aws creds, get these from team lead
-- Replace `NOTIFICATION_QUEUE_PREFIX` with `local_dev_<your org>_`
 - Replace `NOTIFY_EMAIL_DOMAIN` with the domain your emails will come from (i.e. the "origination email" in your SES project)
 - Replace `SECRET_KEY` and `DANGEROUS_SALT` with high-entropy secret values
 - Set up AWS SES and SNS as indicated in next section (AWS Setup), fill in missing AWS env vars

app/celery/tasks.py

@@ -449,6 +449,7 @@ def handle_exception(task, notification, notification_id, exc):
         # Sometimes, SQS plays the same message twice. We should be able to catch an IntegrityError, but it seems
         # SQLAlchemy is throwing a FlushError. So we check if the notification id already exists then do not
         # send to the retry queue.
+        # This probably (hopefully) is not an issue with Redis as the celery backing store
         current_app.logger.exception('Retry' + retry_msg)
         try:
             task.retry(queue=QueueNames.RETRY, exc=exc)

app/config.py

@@ -114,9 +114,6 @@ class Config(object):
     FIRETEXT_API_KEY = os.environ.get("FIRETEXT_API_KEY", "placeholder")
     FIRETEXT_INTERNATIONAL_API_KEY = os.environ.get("FIRETEXT_INTERNATIONAL_API_KEY", "placeholder")
 
-    # Prefix to identify queues in SQS
-    NOTIFICATION_QUEUE_PREFIX = os.environ.get('NOTIFICATION_QUEUE_PREFIX')
-
     # Use notify.sandbox.10x sending domain unless overwritten by environment
     NOTIFY_EMAIL_DOMAIN = 'notify.sandbox.10x.gsa.gov'
 
@@ -200,11 +197,9 @@ class Config(object):
     DVLA_EMAIL_ADDRESSES = json.loads(os.environ.get('DVLA_EMAIL_ADDRESSES', '[]'))
 
     CELERY = {
-        'broker_url': 'sqs://',
+        'broker_url': REDIS_URL,
         'broker_transport_options': {
-            'region': AWS_REGION,
             'visibility_timeout': 310,
-            'queue_name_prefix': NOTIFICATION_QUEUE_PREFIX,
         },
         'timezone': 'Europe/London',
         'imports': [
@@ -447,7 +442,6 @@ class Development(Config):
     NOTIFY_EMAIL_DOMAIN = os.getenv('NOTIFY_EMAIL_DOMAIN', 'notify.sandbox.10x.gsa.gov')
 
     SQLALCHEMY_DATABASE_URI = os.environ.get('SQLALCHEMY_DATABASE_URI', 'postgresql://postgres:chummy@db:5432/notification_api')
-    REDIS_URL = os.environ.get('REDIS_URL')
 
     ANTIVIRUS_ENABLED = os.environ.get('ANTIVIRUS_ENABLED') == '1'
 
@@ -572,7 +566,6 @@ class Live(Config):
     REDIS_ENABLED = os.environ.get('REDIS_ENABLED')
 
     NOTIFY_LOG_PATH = os.environ.get('NOTIFY_LOG_PATH', 'application.log')
-    REDIS_URL = os.environ.get('REDIS_URL')
 
 
 class CloudFoundryConfig(Config):

app/letters/rest.py

@@ -10,6 +10,7 @@ letter_job = Blueprint("letter-job", __name__)
 register_errors(letter_job)
 
 # too many references will make SQS error (as the task can only be 256kb)
+# Maybe doesn't matter anymore with Redis as the celery backing store
 MAX_REFERENCES_PER_TASK = 5000
 
 

app/v2/notifications/post_notifications.py

@@ -236,6 +236,7 @@ def process_sms_or_email_notification(
             # If SQS cannot put the task on the queue, it's probably because the notification body was too long and it
             # went over SQS's 256kb message limit. If the body is very large, it may exceed the HTTP max content length;
             # the exception we get here isn't handled correctly by botocore - we get a ResponseParserError instead.
+            # Hopefully this is no longer an issue with Redis as celery's backing store
             current_app.logger.info(
                 f'Notification {notification_id} failed to save to high volume queue. Using normal flow instead'
             )

manifest.yml

@@ -27,7 +27,6 @@ applications:
       NOTIFY_ENVIRONMENT: live
       API_HOST_NAME: https://notifications-api.app.cloud.gov
       ADMIN_BASE_URL: https://notifications-admin.app.cloud.gov
-      NOTIFICATION_QUEUE_PREFIX: prototype_10x
       STATSD_HOST: localhost
 
       # Credentials variables

pytest.ini

@@ -4,7 +4,6 @@ env =
     NOTIFY_ENVIRONMENT=test
     MMG_API_KEY=mmg-secret-key
     FIRETEXT_API_KEY=Firetext
-    NOTIFICATION_QUEUE_PREFIX=testing
     REDIS_ENABLED=0
 addopts = -p no:warnings
 xfail_strict = true

requirements.in

@@ -2,7 +2,7 @@
 # with package version changes made in requirements.in
 
 cffi==1.15.0
-celery[sqs]==5.2.6
+celery[redis]==5.2.7
 Flask-Bcrypt==1.0.1
 flask-marshmallow==0.14.0
 Flask-Migrate==3.1.0

requirements.txt

@@ -39,7 +39,7 @@ cachetools==5.1.0
     # via
     #   -r requirements.in
     #   notifications-utils
-celery[sqs]==5.2.6
+celery[redis]==5.2.7
     # via -r requirements.in
 certifi==2022.5.18.1
     # via
@@ -219,7 +219,9 @@ pyyaml==5.4.1
     #   awscli
     #   notifications-utils
 redis==4.3.1
-    # via flask-redis
+    # via
+    #   celery
+    #   flask-redis
 requests==2.27.1
     # via
     #   awscli-cwlogs

sample.env

@@ -4,7 +4,6 @@
 DEBUG=True
 ANTIVIRUS_ENABLED=0
 NOTIFY_ENVIRONMENT=development
-NOTIFICATION_QUEUE_PREFIX=local_dev_YOURNAME_
 STATSD_HOST=localhost
 SES_STUB_URL=None
 NOTIFY_APP_NAME=api
@@ -47,4 +46,4 @@ AWS_REGION=us-west-2
 AWS_ACCESS_KEY_ID="don't write secrets to the sample file"
 AWS_SECRET_ACCESS_KEY="don't write secrets to the sample file"
 AWS_PINPOINT_REGION=us-west-2
-AWS_US_TOLL_FREE_NUMBER=+18446120782
+AWS_US_TOLL_FREE_NUMBER=+18446120782

tests/app/service/send_notification/test_send_notification.py

@@ -683,7 +683,7 @@ def test_should_persist_notification(
     (SMS_TYPE, 'send-sms-tasks'),
     (EMAIL_TYPE, 'send-email-tasks')
 ])
-def test_should_delete_notification_and_return_error_if_sqs_fails(
+def test_should_delete_notification_and_return_error_if_redis_fails(
     client,
     sample_email_template,
     sample_template,
@@ -694,7 +694,7 @@ def test_should_delete_notification_and_return_error_if_sqs_fails(
 ):
     mocked = mocker.patch(
         'app.celery.provider_tasks.deliver_{}.apply_async'.format(template_type),
-        side_effect=Exception("failed to talk to SQS")
+        side_effect=Exception("failed to talk to redis")
     )
     mocker.patch('app.notifications.process_notifications.uuid.uuid4', return_value=fake_uuid)
 
@@ -719,7 +719,7 @@ def test_should_delete_notification_and_return_error_if_sqs_fails(
             data=json.dumps(data),
             headers=[('Content-Type', 'application/json'), ('Authorization', 'Bearer {}'.format(auth_header))]
         )
-    assert str(e.value) == 'failed to talk to SQS'
+    assert str(e.value) == 'failed to talk to redis'
 
     mocked.assert_called_once_with([fake_uuid], queue=queue_name)
     assert not notifications_dao.get_notification_by_id(fake_uuid)