celery_tasks

#!/usr/bin/env python

"""=cut
=head1 NAME

celery_tasks - Munin plugin to monitor the number of Celery tasks with specified names.

=head1 REQUIREMENTS

 - Python
 - celery (http://celeryproject.org/)
 - celerymon (http://github.com/ask/celerymon)

Note: don't forget to enable sending of the events on the celery daemon - run it with the --events option

=head1 CONFIGURATION

Default configuration:

None

You must set the name of at least one task you want to monitor (multiple names are separated by a comma).

For example:

  [celery_tasks]
	 env.tasks myapp.tasks.SendEmailTask,myapp2.tasks.FetchUserDataTask

This would monitor the number of task for a task with name "myapp.tasks.SendEmailTask" and "myapp2.tasks.FetchUserDataTask".

=head1 MAGIC MARKERS

  #%# family=manual
  #%# capabilities=autoconf

=head1 AUTHOR

Tomaz Muraus (http://github.com/Kami/munin-celery)

=head1 LICENSE

GPLv2

=cut"""

import os
import sys
import urllib

try:
	import json
except:
	import simplejson as json

API_URL = 'http://localhost:8989'
URL_ENDPOINTS = {
		'workers': '/api/worker/',
		'worker_tasks': '/api/worker/%s/tasks',
		'tasks': '/api/task/',
		'task_names': '/api/task/name/',
		'task_details': '/api/task/name/%s',
}
TASK_STATES = (
			'PENDING',
			'RECEIVED',
			'STARTED',
			'SUCCESS',
			'FAILURE',
			'REVOKED',
			'RETRY'
)

def get_data(what, api_url, *args):
	try:
		request = urllib.urlopen('%s%s' % (api_url, \
										   URL_ENDPOINTS[what] % (args)))
		response = request.read()
		return json.loads(response)
	except IOError:
		print 'Could not connect to the celerymon webserver'
		sys.exit(-1)

def check_web_server_status(api_url):
	try:
		request = urllib.urlopen(api_url)
		response = request.read()
	except IOError:
		print 'Could not connect to the celerymon webserver'
		sys.exit(-1)

def clean_task_name(task_name):
	return task_name.replace('.', '_')

# Config
def print_config(task_names):
	print 'graph_title Celery tasks'
	print 'graph_args --lower-limit 0'
	print 'graph_scale no'
	print 'graph_vlabel tasks per ${graph_period}'
	print 'graph_category cloud'

	for name in task_names:
		print '%s.label %s' % (clean_task_name(name), name)
		print '%s.type DERIVE' % (clean_task_name(name))
		print '%s.min 0' % (clean_task_name(name))
		print '%s.info number of %s tasks' % (clean_task_name(name), name)

# Values
def print_values(task_names = None, api_url = None):
	for task_name in task_names:
		count = len(get_data('task_details', api_url, task_name))
		print '%s.value %d' % (clean_task_name(task_name), count)

if __name__ == '__main__':
	task_names = os.environ.get('tasks', None)
	api_url = os.environ.get('api_url', API_URL)

	check_web_server_status(api_url)

	if not task_names:
		print 'You need to define at least one task name'
		sys.exit(-1)

	task_names = task_names.split(',')

	if len(sys.argv) > 1:
		if sys.argv[1] == 'config':
			print_config(task_names)
		elif sys.argv[1] == 'autoconf':
			print 'yes'
	else:
		print_values(task_names, api_url)