Celery CloudWatch

Monitor your celery application from within AWS CloudWatch!

Metrics

The following events are tallied per task:

• CeleryEventSent
• CeleryEventStarted
• CeleryEventSucceeded
• CeleryEventFailed

You can then see how many tasks/day, tasks/week etc are being completed.

Also, we record the total waiting/running per task:

• CeleryNumWaiting
• CeleryNumRunning

Also, statistics on task duration are sent in the metrics:

• CeleryWaitingTime
• CeleryProcessingTime

These metrics are sent with all supported stats (No. Events, Sum, Max, Min), allowing you to gain insight into your task processing and match requests and capacity.

Getting Started

1. Set up an IAM Role for your instance.

   It must include a policy to perform 'PutMetricData', eg:

   {
     "Version": "2000-01-01",
     "Statement": [
       {
         "Effect": "Allow",
         "Action": [
           "cloudwatch:PutMetricData"
         ],
         "Resource": [
           "*"
         ]
       }
     ]
   }
   

   (Note: Alternatively, you can set up a User with the same policy and provide access details that way)

2. Install via python-pip (and upgrade pip & boto)

   sudo apt-get install -y python-pip
   sudo pip install --upgrade pip boto
   
   # Install directly
   sudo pip install celery-cloudwatch
   
   # OR, install in a virtualenv
   sudo apt-get install -y python-virtualenv
   mkdir /var/python-envs
   virtualenv /var/python-envs/ccwatch
   source /var/python-envs/ccwatch/bin/activate
   pip install celery-cloudwatch

3. Create your own boto.cfg at /etc/boto.cfg-

   [Credentials]
   # if not using an IAM Role - provide aws key/secret
   aws_access_key_id = xxx
   aws_secret_access_key = yyy
   
   [Boto]
   cloudwatch_region_name = my-region
   cloudwatch_region_endpoint = monitoring.my-region.amazonaws.com
   
   

4. Create your own config file in /etc/ccwatch.yaml

   ccwatch:
     broker: null
     camera: celery_cloudwatch.CloudWatchCamera
     verbose: no
   camera:
     frequency: 60.0
     verbose: no
   cloudwatch-camera:
     dryrun: no
     namespace: celery
     tasks:
       - myapp.mytasks.taskname
       - myapp.mytasks.anothertask
       - myapp.mytasks.thirdtask
       - name: myapp.secondarytasks
         dimensions:
           task: myapp.secondarytasks
           customDim: value
       - name: myapp.tertiarytasks
         dimensions:
           task: myapp.tertiarytasks
           customDim: value
   
   

5. Install upstart

   Create a file /etc/init/celery-cloudwatch.conf-

   description "Celery CloudWatch"
   author "nathan muir <ndmuir@gmail.com>"
   
   setuid nobody
   setgid nogroup
   
   start on runlevel [234]
   stop on runlevel [0156]
   
   exec /var/python-envs/ccwatch/bin/ccwatch
   respawn
   

   then-

   sudo initctl reload-configuration
   sudo service celery-cloudwatch start

6. Start Celery your celery workers with the -E (or CELERY_SEND_EVENTS=1 and CELERY_TRACK_STARTED=1) options, and, start celery clients with CELERY_SEND_TASK_SENT_EVENT=1

7. All done! head over to your CloudWatch monitoring page to see the results!