NSQ Nagios Scripts

check_aggregate_nsqd_depth.py

#!/usr/bin/env python2.7
import sys
import logging
import functools
import time
import tornado.options
import tornado.ioloop
import nsq_data


OK = 0 # The plugin was able to check the service and it appeared to be functioning properly
WARNING = 1 # The plugin was able to check the service, but it appeared to be above some "warning" threshold or did not appear to be working properly
CRITICAL = 2 # The plugin detected that either the service was not running or it was above some "critical" threshold
UNKNOWN = 3 # Invalid command line arguments were supplied to the plugin or low-level failures internal to the plugin

state = UNKNOWN
message = "failed"
cluster_data = {}
num_outstanding_requests = 0


def get_topics(endpoints):
    global num_outstanding_requests
    
    num_outstanding_requests = len(endpoints)
    for endpoint in endpoints:
        callback = functools.partial(_finish_get_topics, endpoints=endpoints)
        nsq_data.get_topics_in_lookupd(endpoint, callback)

def _finish_get_topics(lookup_data, endpoints):
    global cluster_data
    global num_outstanding_requests
    
    num_outstanding_requests -= 1
    
    lookup_data = lookup_data or {}
    for topic in lookup_data.get('topics', []):
        cluster_data[topic] = {}
    
    if num_outstanding_requests == 0:
        get_producers(endpoints)


def get_producers(endpoints):
    global num_outstanding_requests
    
    for topic in cluster_data:
        for endpoint in endpoints:
            num_outstanding_requests += 1
            nsq_data.get_producers_for_topic(endpoint, topic, functools.partial(_finish_get_producers, topic=topic))

def _finish_get_producers(lookup_data, topic):
    global cluster_data
    global num_outstanding_requests
    
    num_outstanding_requests -= 1
    
    lookup_data = lookup_data or {}
    for channel in lookup_data.get('channels', []):
        if channel not in cluster_data[topic] and not channel.endswith('#ephemeral'):
            cluster_data[topic][channel] = {}
            # None means no producers
            cluster_data[topic][channel]['depth'] = None
            cluster_data[topic][channel]['producers'] = set()
    
    for producer in lookup_data.get('producers', []):
        producer_endpoint = 'http://%s:%d' % (producer['broadcast_address'], producer['http_port'])
        for channel in cluster_data[topic]:
            cluster_data[topic][channel]['producers'].add(producer_endpoint)
    
    if num_outstanding_requests == 0:
        get_depths()


def get_depths():
    global num_outstanding_requests
    
    endpoints = set()
    for topic in cluster_data:
        for channel in cluster_data[topic]:
            # we have producers, set depth to 0
            if cluster_data[topic][channel]['producers']:
                cluster_data[topic][channel]['depth'] = 0
            for endpoint in cluster_data[topic][channel]['producers']:
                endpoints.add(endpoint)
    
    if not endpoints:
        finish()
    
    for endpoint in endpoints:
        num_outstanding_requests += 1
        nsq_data.get_nsqd_stats(endpoint, functools.partial(_finish_get_depths, topic=topic, channel=channel))

def _finish_get_depths(nsqd_data, topic, channel):
    global num_outstanding_requests
    global cluster_data
    
    num_outstanding_requests -= 1
    
    nsqd_data = nsqd_data or {}
    for topic_data in nsqd_data.get('topics', []):
        topic = topic_data['topic_name']
        for channel_data in topic_data.get('channels', []):
            channel = channel_data['channel_name']
            if channel in cluster_data.get(topic, {}):
                cluster_data[topic][channel]['depth'] += channel_data.get('depth', 0) + channel_data.get('deferred_count', 0)
    
    if num_outstanding_requests == 0:
        finish()

def finish():
    global state
    global message
    
    topic = tornado.options.options.topic
    channel = tornado.options.options.channel
    warning_threshold = tornado.options.options.warning
    critical_threshold = tornado.options.options.critical
    
    channel_data = cluster_data.get(topic, {}).get(channel, {})
    depth = channel_data.get('depth')
    
    if not channel_data or depth is None:
        state = 2
        message = 'no producers found in lookupd'
    else:
        state = 0
        message = 'depth of %d' % depth
    
    if depth > warning_threshold:
        state = 1
        message = 'depth of %d (warn at %d)' % (depth, warning_threshold)
    if depth > critical_threshold:
        state = 2
        message = 'depth of %d (crit at %d)' % (depth, critical_threshold)
    
    tornado.ioloop.IOLoop.instance().stop()


def main():
    tornado.options.define("lookupd_http_address", default="http://127.0.0.1:4161", multiple=True,
        help="http address of the lookupd instances in this cluster (multiple)")
    tornado.options.define("warning", default=50000, help="warning threshold")
    tornado.options.define("critical", default=100000, help="critical threshold")
    tornado.options.define("topic", help="NSQ topic")
    tornado.options.define("channel", help="NSQ channel")
    tornado.options.parse_command_line()
    
    logging.getLogger().setLevel(logging.CRITICAL)
    
    get_topics(tornado.options.options.lookupd_http_address)
    
    io_loop = tornado.ioloop.IOLoop.instance()
    # add a deadline to prevent it from living forever in exceptional states
    io_loop.add_timeout(time.time() + 9, io_loop.stop)
    io_loop.start()
    
    print message
    return state


if __name__ == "__main__":
    sys.exit(main())

check_nsqd_registrations.py

#!/usr/bin/env python2.7
import sys
import tornado.ioloop
import tornado.options
import logging
import functools
import time
import nsq_data


OK = 0 # The plugin was able to check the service and it appeared to be functioning properly
WARNING = 1 # The plugin was able to check the service, but it appeared to be above some "warning" threshold or did not appear to be working properly
CRITICAL = 2 # The plugin detected that either the service was not running or it was above some "critical" threshold
UNKNOWN = 3 # Invalid command line arguments were supplied to the plugin or low-level failures internal to the plugin

state = UNKNOWN
message = "failed"
cluster_data = {}
num_outstanding_requests = 0


def get_topics(endpoints):
    global num_outstanding_requests
    
    num_outstanding_requests = len(endpoints)
    for endpoint in endpoints:
        callback = functools.partial(_finish_get_topics, endpoints=endpoints)
        nsq_data.get_topics_in_lookupd(endpoint, callback)

def _finish_get_topics(lookup_data, endpoints):
    global cluster_data
    global num_outstanding_requests
    
    num_outstanding_requests -= 1
    
    lookup_data = lookup_data or {}
    for topic in lookup_data.get('topics', []):
        cluster_data[topic] = {}
    
    if num_outstanding_requests == 0:
        get_producers(endpoints)


def get_producers(endpoints):
    global num_outstanding_requests
    
    for topic in cluster_data:
        for endpoint in endpoints:
            num_outstanding_requests += 1
            nsq_data.get_producers_for_topic(endpoint, topic, functools.partial(_finish_get_producers, topic=topic))

def _finish_get_producers(lookup_data, topic):
    global cluster_data
    global num_outstanding_requests
    
    num_outstanding_requests -= 1
    
    lookup_data = lookup_data or {}
    for channel in lookup_data.get('channels', []):
        if channel not in cluster_data[topic] and not channel.endswith('#ephemeral'):
            cluster_data[topic][channel] = {}
            cluster_data[topic][channel]['producers'] = set()
    
    for producer in lookup_data.get('producers', []):
        producer_endpoint = 'http://%s:%d' % (producer['broadcast_address'], producer['http_port'])
        for channel in cluster_data[topic]:
            cluster_data[topic][channel]['producers'].add(producer_endpoint)
    
    if num_outstanding_requests == 0:
        get_stats()


def get_stats():
    endpoint = tornado.options.options.nsqd_http_address
    nsq_data.get_nsqd_stats(endpoint, functools.partial(_finish_get_stats, endpoint=endpoint))

def _finish_get_stats(nsqd_data, endpoint):
    global state
    global message
    
    if nsqd_data is None:
        state = 2
        message = 'failed to get information from nsqd'
    elif not nsqd_data:
        state = 1
        message = 'no topics/channels'
    else:
        missing = []
        for topic_data in nsqd_data.get('topics', []):
            topic = topic_data['topic_name']
            for channel_data in topic_data.get('channels', []):
                channel = channel_data['channel_name']
                producers = cluster_data.get(topic, {}).get(channel, {}).get('producers', [])
                if endpoint not in producers and not channel.endswith('#ephemeral'):
                    missing.append("%s/%s" % (topic, channel))
        if missing:
            state = 2
            message = 'missing - %s' % ', '.join(missing)
        else:
            state = 0
            message = 'all topics/channels registered in cluster'
    
    tornado.ioloop.IOLoop.instance().stop()


def main():
    tornado.options.define("lookupd_http_address", default="http://127.0.0.1:4161", multiple=True,
        help="http address of the lookupd instances in this cluster (multiple)")
    tornado.options.define("warning", default=50000, help="warning threshold")
    tornado.options.define("critical", default=100000, help="critical threshold")
    tornado.options.define("nsqd_http_address", default="http://127.0.0.1:4151", 
        help="http address of the nsqd instance to check")
    tornado.options.parse_command_line()
    
    logging.getLogger().setLevel(logging.CRITICAL)
    
    get_topics(tornado.options.options.lookupd_http_address)
    
    io_loop = tornado.ioloop.IOLoop.instance()
    # add a deadline to prevent it from living forever in exceptional states
    io_loop.add_timeout(time.time() + 9, io_loop.stop)
    io_loop.start()
    
    print message
    return state


if __name__ == '__main__':
    sys.exit(main())

nsq_data.py

import tornado.httpclient
import logging
import json
import urllib
import functools


_http_client = None
def http_client():
    global _http_client
    if not _http_client:
        _http_client = tornado.httpclient.AsyncHTTPClient()
    return _http_client

def _process_response(response, endpoint, user_callback):
    if response.error:
        logging.warning("[%s] http error %s", endpoint, response.error)
        return user_callback(None)
    
    try:
        data = json.loads(response.body)
    except json.JSONDecodeError:
        logging.warning("[%s] failed to parse JSON: %r", endpoint, response.body)
        return user_callback(None)
    
    if data.get('status_code') != 200:
        logging.warning("[%s] endpoint responded with %d", endpoint, data.get('status_code'))
        return user_callback(None)
    
    user_callback(data.get('data', {}))

def get_topics_in_lookupd(endpoint, user_callback):
    lookupd_url = endpoint + "/topics"
    req = tornado.httpclient.HTTPRequest(lookupd_url, method="GET", 
        connect_timeout=2, request_timeout=5)
    callback = functools.partial(_process_response, endpoint=endpoint, user_callback=user_callback)
    http_client().fetch(req, callback=callback)

def get_producers_for_topic(endpoint, topic, user_callback):
    lookupd_url = endpoint + "/lookup?topic=" + urllib.quote(topic)
    req = tornado.httpclient.HTTPRequest(lookupd_url, method="GET", 
        connect_timeout=2, request_timeout=5)
    callback = functools.partial(_process_response, endpoint=endpoint, user_callback=user_callback)
    http_client().fetch(req, callback=callback)

def get_nsqd_stats(endpoint, user_callback):
    nsqd_url = endpoint + "/stats?format=json"
    req = tornado.httpclient.HTTPRequest(nsqd_url, method="GET", 
        connect_timeout=2, request_timeout=5)
    callback = functools.partial(_process_response, endpoint=endpoint, user_callback=user_callback)
    http_client().fetch(req, callback=callback)