Merge pull request #3201 from alphagov/revamp-celery-stats

Migrate towards new metrics for Celery tasks

app/__init__.py

@@ -110,7 +110,7 @@ def create_app(application):
     email_clients = [aws_ses_stub_client] if application.config['SES_STUB_URL'] else [aws_ses_client]
     notification_provider_clients.init_app(sms_clients=[firetext_client, mmg_client], email_clients=email_clients)
 
-    notify_celery.init_app(application)
+    notify_celery.init_app(application, statsd_client)
     encryption.init_app(application)
     redis_store.init_app(application)
     document_download_client.init_app(application)

app/celery/celery.py

@@ -4,7 +4,6 @@ from celery import Celery, Task
 from celery.signals import worker_process_shutdown
 from flask import g, request
 from flask.ctx import has_app_context, has_request_context
-from gds_metrics.metrics import Histogram
 
 
 @worker_process_shutdown.connect
@@ -19,60 +18,82 @@ def log_on_worker_shutdown(sender, signal, pid, exitcode, **kwargs):
         notify_celery._app.logger.info('worker shutdown: PID: {} Exitcode: {}'.format(pid, exitcode))
 
 
-def make_task(app):
-    SQS_APPLY_ASYNC_DURATION_SECONDS = Histogram(
-        'sqs_apply_async_duration_seconds',
-        'Time taken to put task on queue',
-        ['task_name']
-    )
-
+def make_task(app, statsd_client):
     class NotifyTask(Task):
         abstract = True
         start = None
 
         def on_success(self, retval, task_id, args, kwargs):
-            elapsed_time = time.time() - self.start
+            elapsed_time = time.monotonic() - self.start
+            delivery_info = self.request.delivery_info or {}
+            queue_name = delivery_info.get('routing_key', 'none')
+
             app.logger.info(
-                "{task_name} took {time}".format(
-                    task_name=self.name, time="{0:.4f}".format(elapsed_time)
+                "Celery task {task_name} (queue: {queue_name}) took {time}".format(
+                    task_name=self.name,
+                    queue_name=queue_name,
+                    time="{0:.4f}".format(elapsed_time)
                 )
             )
 
+            statsd_client.timing(
+                "celery.{queue_name}.{task_name}.success".format(
+                    task_name=self.name,
+                    queue_name=queue_name
+                ), elapsed_time
+            )
+
         def on_failure(self, exc, task_id, args, kwargs, einfo):
-            # ensure task will log exceptions to correct handlers
-            app.logger.exception('Celery task: {} failed'.format(self.name))
+            delivery_info = self.request.delivery_info or {}
+            queue_name = delivery_info.get('routing_key', 'none')
+
+            app.logger.exception(
+                "Celery task {task_name} (queue: {queue_name}) failed".format(
+                    task_name=self.name,
+                    queue_name=queue_name,
+                )
+            )
+
+            statsd_client.incr(
+                "celery.{queue_name}.{task_name}.failure".format(
+                    task_name=self.name,
+                    queue_name=queue_name
+                )
+            )
+
             super().on_failure(exc, task_id, args, kwargs, einfo)
 
         def __call__(self, *args, **kwargs):
             # ensure task has flask context to access config, logger, etc
             with app.app_context():
-                self.start = time.time()
-                # Remove 'request_id' from the kwargs (so the task doesn't get an unexpected kwarg), then add it to g
-                # so that it gets logged
+                self.start = time.monotonic()
+                # Remove piggyback values from kwargs
+                # Add 'request_id' to 'g' so that it gets logged
                 g.request_id = kwargs.pop('request_id', None)
+
                 return super().__call__(*args, **kwargs)
 
         def apply_async(self, args=None, kwargs=None, task_id=None, producer=None,
                         link=None, link_error=None, **options):
             kwargs = kwargs or {}
+
             if has_request_context() and hasattr(request, 'request_id'):
                 kwargs['request_id'] = request.request_id
             elif has_app_context() and 'request_id' in g:
                 kwargs['request_id'] = g.request_id
 
-            with SQS_APPLY_ASYNC_DURATION_SECONDS.labels(self.name).time():
-                return super().apply_async(args, kwargs, task_id, producer, link, link_error, **options)
+            return super().apply_async(args, kwargs, task_id, producer, link, link_error, **options)
 
     return NotifyTask
 
 
 class NotifyCelery(Celery):
 
-    def init_app(self, app):
+    def init_app(self, app, statsd_client):
         super().__init__(
             app.import_name,
             broker=app.config['BROKER_URL'],
-            task_cls=make_task(app),
+            task_cls=make_task(app, statsd_client),
         )
 
         self.conf.update(app.config)