def run(self):
self.running = True
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=LoggerMessage(self.mq))
self.mq.subscribe(destination=CONF.outbound_queue)
while not self.shuttingdown:
try:
LOG.debug('Waiting for log messages...')
time.sleep(CONF.loop_every)
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
def run(self):
self.running = True
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=GangliaMessage(self.mq))
while not self.shuttingdown:
try:
rules = init_rules() # re-read rule config each time
self.metric_check(rules)
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
LOG.debug('Waiting for next check run...')
time.sleep(CONF.loop_every)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
def run(self):
self.running = True
# Create internal queue
self.queue = Queue.Queue()
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=UrlmonMessage(self.mq))
self.dedup = DeDup()
self.carbon = Carbon() # graphite metrics
# Initialiase alert rules
urls = init_urls()
# Start worker threads
LOG.debug('Starting %s worker threads...', CONF.server_threads)
for i in range(CONF.server_threads):
w = WorkerThread(self.mq, self.queue, self.dedup, self.carbon)
try:
w.start()
except Exception, e:
LOG.error('Worker thread #%s did not start: %s', i, e)
continue
LOG.info('Started worker thread: %s', w.getName())
class AlertaDaemon(Daemon):
alerta_opts = {
'forward_duplicate': 'no',
}
def __init__(self, prog, **kwargs):
config.register_opts(AlertaDaemon.alerta_opts)
Daemon.__init__(self, prog, kwargs)
def run(self):
self.running = True
self.queue = Queue.Queue() # Create internal queue
self.db = Mongo() # mongo database
self.carbon = Carbon() # carbon metrics
self.statsd = StatsD() # graphite metrics
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=ServerMessage(self.mq, self.queue, self.statsd))
self.mq.subscribe()
# Start worker threads
LOG.debug('Starting %s worker threads...', CONF.server_threads)
for i in range(CONF.server_threads):
w = WorkerThread(self.mq, self.queue, self.statsd)
try:
w.start()
except Exception, e:
LOG.error('Worker thread #%s did not start: %s', i, e)
continue
LOG.info('Started worker thread: %s', w.getName())
while not self.shuttingdown:
try:
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version, timeout=CONF.loop_every)
self.mq.send(heartbeat)
time.sleep(CONF.loop_every)
LOG.info('Alert processing queue length is %d', self.queue.qsize())
self.carbon.metric_send('alerta.alerts.queueLength', self.queue.qsize())
self.db.update_queue_metric(self.queue.qsize())
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
for i in range(CONF.server_threads):
self.queue.put(None)
w.join()
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
def run(self):
self.running = True
self.queue = Queue.Queue() # Create internal queue
self.db = Mongo() # mongo database
self.carbon = Carbon() # carbon metrics
self.statsd = StatsD() # graphite metrics
# Connect to message queue
self.mq = Messaging()
self.mq.connect(
callback=ServerMessage(self.mq, self.queue, self.statsd))
self.mq.subscribe()
# Start worker threads
LOG.debug('Starting %s worker threads...', CONF.server_threads)
for i in range(CONF.server_threads):
w = WorkerThread(self.mq, self.queue, self.statsd)
try:
w.start()
except Exception, e:
LOG.error('Worker thread #%s did not start: %s', i, e)
continue
LOG.info('Started worker thread: %s', w.getName())
def run(self):
self.running = True
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=DynectMessage(self.mq))
while not self.shuttingdown:
try:
self.queryDynect()
if self.updating:
self.alertDynect()
self.last_info = self.info
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
LOG.debug('Waiting for next check run...')
time.sleep(CONF.loop_every)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
self.running = False
def run(self):
self.running = True
# Start token bucket thread
self.tokens = LeakyBucket(tokens=20, rate=30)
self.tokens.start()
self.onhold = dict()
# Connect to message queue
self.mq = Messaging()
self.mq.connect(
callback=MailerMessage(self.mq, self.onhold, self.tokens))
self.mq.subscribe(destination=CONF.outbound_topic)
while not self.shuttingdown:
try:
LOG.debug('Send email messages...')
for alertid in self.onhold.keys():
try:
(mailAlert, hold_time) = self.onhold[alertid]
except KeyError:
continue
if time.time() > hold_time:
if not self.tokens.get_token():
LOG.warning(
'%s : No tokens left, rate limiting this alert',
alertid)
continue
email = Mailer(mailAlert)
mail_to = CONF.mail_list.split(',')
if 'mailto' in mailAlert.tags:
mail_to.append(mailAlert.tags['mailto'])
email.send(mail_to=mail_to)
try:
del self.onhold[alertid]
except KeyError:
continue
time.sleep(CONF.loop_every)
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
self.tokens.shutdown()
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
class UrlmonDaemon(Daemon):
def run(self):
self.running = True
# Create internal queue
self.queue = Queue.Queue()
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=UrlmonMessage(self.mq))
self.dedup = DeDup()
self.statsd = StatsD() # graphite metrics
# Initialiase alert rules
urls = init_urls()
# Start worker threads
LOG.debug('Starting %s worker threads...', CONF.server_threads)
for i in range(CONF.server_threads):
w = WorkerThread(self.mq, self.queue, self.dedup, self.statsd)
try:
w.start()
except Exception, e:
LOG.error('Worker thread #%s did not start: %s', i, e)
continue
LOG.info('Started worker thread: %s', w.getName())
while not self.shuttingdown:
try:
for url in urls:
self.queue.put(url)
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
LOG.info('URL check queue length is %d', self.queue.qsize())
time.sleep(CONF.loop_every)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
for i in range(CONF.server_threads):
self.queue.put(None)
w.join()
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
class PagerDutyDaemon(Daemon):
def run(self):
self.running = True
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=PagerDutyMessage(self.mq))
self.mq.subscribe(destination=CONF.outbound_topic) # TODO(nsatterl): use dedicated queue?
while not self.shuttingdown:
try:
LOG.debug('Waiting for PagerDuty messages...')
time.sleep(CONF.loop_every)
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
class LoggerDaemon(Daemon):
"""
Index alerts in ElasticSearch using Logstash format so that logstash GUI and/or Kibana can be used as front-ends
"""
def run(self):
self.running = True
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=LoggerMessage())
self.mq.subscribe(destination=CONF.outbound_queue)
while not self.shuttingdown:
try:
LOG.debug('Waiting for log messages...')
time.sleep(30)
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat()
self.mq.send(heartbeat)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
class MailerDaemon(Daemon):
def run(self):
self.running = True
# Start token bucket thread
self.tokens = LeakyBucket(tokens=20, rate=30)
self.tokens.start()
self.onhold = dict()
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=MailerMessage(self.mq, self.onhold, self.tokens))
self.mq.subscribe(destination=CONF.outbound_topic)
while not self.shuttingdown:
try:
LOG.debug('Send email messages...')
for alertid in self.onhold.keys():
try:
(mailAlert, hold_time) = self.onhold[alertid]
except KeyError:
continue
if time.time() > hold_time:
if not self.tokens.get_token():
LOG.warning('%s : No tokens left, rate limiting this alert', alertid)
continue
email = Mailer(mailAlert)
mail_to = CONF.mail_list.split(',')
for tag in mailAlert.tags:
if tag.startswith('email'):
mail_to.append(tag.split(':')[1])
email.send(mail_to=mail_to)
try:
del self.onhold[alertid]
except KeyError:
continue
time.sleep(CONF.loop_every)
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
self.tokens.shutdown()
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
def run(self):
self.running = True
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=LoggerMessage(self.mq))
self.mq.subscribe(destination=CONF.outbound_queue)
while not self.shuttingdown:
try:
LOG.debug('Waiting for log messages...')
time.sleep(CONF.loop_every)
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
def run(self):
self.running = True
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=DynectMessage(self.mq))
while not self.shuttingdown:
try:
self.queryDynect()
if self.updating:
self.alertDynect()
self.last_info = self.info
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
LOG.debug('Waiting for next check run...')
time.sleep(CONF.loop_every)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
self.running = False
def run(self):
self.running = True
# Create internal queue
self.queue = Queue.Queue()
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=UrlmonMessage(self.mq))
self.dedup = DeDup()
self.carbon = Carbon() # graphite metrics
# Initialiase alert rules
urls = init_urls()
# Start worker threads
LOG.debug('Starting %s worker threads...', CONF.server_threads)
for i in range(CONF.server_threads):
w = WorkerThread(self.mq, self.queue, self.dedup, self.carbon)
try:
w.start()
except Exception, e:
LOG.error('Worker thread #%s did not start: %s', i, e)
continue
LOG.info('Started worker thread: %s', w.getName())
def run(self):
self.running = True
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=GangliaMessage(self.mq))
while not self.shuttingdown:
try:
rules = init_rules() # re-read rule config each time
self.metric_check(rules)
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
LOG.debug('Waiting for next check run...')
time.sleep(CONF.loop_every)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
class AlertaDaemon(Daemon):
def run(self):
self.running = True
self.queue = Queue.Queue() # Create internal queue
self.statsd = StatsD() # graphite metrics
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=ServerMessage(self.mq, self.queue, self.statsd))
self.mq.subscribe()
# Start worker threads
LOG.debug('Starting %s worker threads...', CONF.server_threads)
for i in range(CONF.server_threads):
w = WorkerThread(self.mq, self.queue, self.statsd)
try:
w.start()
except Exception, e:
LOG.error('Worker thread #%s did not start: %s', i, e)
continue
LOG.info('Started worker thread: %s', w.getName())
while not self.shuttingdown:
try:
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version, timeout=CONF.loop_every)
self.mq.send(heartbeat)
LOG.debug('Internal queue size is %s messages', self.queue.qsize())
time.sleep(CONF.loop_every)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
for i in range(CONF.server_threads):
self.queue.put(None)
w.join()
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
def run(self):
self.running = True
# Initialiase alert config
init_config()
# Start token bucket thread
_TokenThread = TokenTopUp()
_TokenThread.start()
# Start notify thread
_NotifyThread = ReleaseThread()
_NotifyThread.start()
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=NotifyMessage(self.mq))
self.mq.subscribe(destination=CONF.outbound_topic)
while not self.shuttingdown:
try:
# Read (or re-read) config as necessary
if os.path.getmtime(CONF.yaml_config) != config_mod_time:
init_config()
config_mod_time = os.path.getmtime(CONF.yaml_config)
LOG.debug('Waiting for email messages...')
time.sleep(CONF.loop_every)
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
_TokenThread.shutdown()
_NotifyThread.shutdown()
LOG.info('Shutdown request received...')
self.running = False
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
class CloudWatchDaemon(Daemon):
cloudwatch_opts = {
'cloudwatch_sqs_region': 'eu-west-1',
'cloudwatch_sqs_queue': 'cloudwatch-to-alerta',
'cloudwatch_access_key': '022QF06E7MXBSAMPLE',
'cloudwatch_secret_key': ''
}
def __init__(self, prog, **kwargs):
config.register_opts(CloudWatchDaemon.cloudwatch_opts)
Daemon.__init__(self, prog, kwargs)
def run(self):
self.running = True
self.statsd = StatsD() # graphite metrics
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=CloudWatchMessage(self.mq))
self.dedup = DeDup(by_value=True)
LOG.info('Connecting to SQS queue %s', CONF.cloudwatch_sqs_queue)
try:
sqs = boto.sqs.connect_to_region(
CONF.cloudwatch_sqs_region,
aws_access_key_id=CONF.cloudwatch_access_key,
aws_secret_access_key=CONF.cloudwatch_secret_key
)
except boto.exception.SQSError, e:
LOG.error('SQS API call failed: %s', e)
sys.exit(1)
try:
q = sqs.create_queue(CONF.cloudwatch_sqs_queue)
q.set_message_class(RawMessage)
except boto.exception.SQSError, e:
LOG.error('SQS queue error: %s', e)
sys.exit(1)
def run(self):
self.running = True
# Connect to message queue
self.mq = Messaging()
self.mq.connect()
# Initialiase alert rules
init_urls()
url_mod_time = os.path.getmtime(URLFILE)
# Start worker threads
for i in range(NUM_THREADS):
w = WorkerThread(queue)
w.start()
LOG.info('Starting thread: %s', w.getName())
while not self.shuttingdown:
try:
# Read (or re-read) urls as necessary
if os.path.getmtime(URLFILE) != url_mod_time:
init_urls()
url_mod_time = os.path.getmtime(URLFILE)
for url in urls:
queue.put(('url', url))
queue.put(('timestamp', time.time()))
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat()
self.mq.send(heartbeat)
time.sleep(_check_rate)
urlmon_qsize = queue.qsize()
LOG.info('URL check queue length is %d', urlmon_qsize)
if GMETRIC_SEND:
gmetric_cmd = "%s --name urlmon_qsize --value %d --type uint16 --units \" \" --slope both --group urlmon %s" % (
GMETRIC_CMD, urlmon_qsize, GMETRIC_OPTIONS)
LOG.debug("%s", gmetric_cmd)
os.system("%s" % gmetric_cmd)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
for i in range(NUM_THREADS):
queue.put(('stop', None))
w.join()
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
class CloudWatchDaemon(Daemon):
cloudwatch_opts = {
'cloudwatch_sqs_region': 'eu-west-1',
'cloudwatch_sqs_queue': 'cloudwatch-to-alerta',
'cloudwatch_access_key': '022QF06E7MXBSAMPLE',
'cloudwatch_secret_key': ''
}
def __init__(self, prog, **kwargs):
config.register_opts(CloudWatchDaemon.cloudwatch_opts)
Daemon.__init__(self, prog, kwargs)
def run(self):
self.running = True
self.statsd = StatsD() # graphite metrics
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=CloudWatchMessage(self.mq))
self.dedup = DeDup(by_value=True)
LOG.info('Connecting to SQS queue %s', CONF.cloudwatch_sqs_queue)
try:
sqs = boto.sqs.connect_to_region(
CONF.cloudwatch_sqs_region,
aws_access_key_id=CONF.cloudwatch_access_key,
aws_secret_access_key=CONF.cloudwatch_secret_key)
except boto.exception.SQSError, e:
LOG.error('SQS API call failed: %s', e)
sys.exit(1)
try:
q = sqs.create_queue(CONF.cloudwatch_sqs_queue)
q.set_message_class(RawMessage)
except boto.exception.SQSError, e:
LOG.error('SQS queue error: %s', e)
sys.exit(1)
def run(self):
self.running = True
self.statsd = StatsD() # graphite metrics
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=CloudWatchMessage(self.mq))
self.dedup = DeDup(by_value=True)
LOG.info('Connecting to SQS queue %s', CONF.cloudwatch_sqs_queue)
try:
sqs = boto.sqs.connect_to_region(
CONF.cloudwatch_sqs_region,
aws_access_key_id=CONF.cloudwatch_access_key,
aws_secret_access_key=CONF.cloudwatch_secret_key)
except boto.exception.SQSError, e:
LOG.error('SQS API call failed: %s', e)
sys.exit(1)
def run(self):
data = sys.stdin.read()
LOG.info('snmptrapd -> %s', data)
snmptrapAlert = self.parse_snmptrap(data)
mq = Messaging()
mq.connect()
mq.send(snmptrapAlert)
mq.disconnect()
class NotifyDaemon(Daemon):
def run(self):
self.running = True
# Initialiase alert config
init_config()
# Start token bucket thread
_TokenThread = TokenTopUp()
_TokenThread.start()
# Start notify thread
_NotifyThread = ReleaseThread()
_NotifyThread.start()
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=NotifyMessage(self.mq))
self.mq.subscribe(destination=CONF.outbound_topic)
while not self.shuttingdown:
try:
# Read (or re-read) config as necessary
if os.path.getmtime(CONF.yaml_config) != config_mod_time:
init_config()
config_mod_time = os.path.getmtime(CONF.yaml_config)
LOG.debug('Waiting for email messages...')
time.sleep(CONF.loop_every)
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
_TokenThread.shutdown()
_NotifyThread.shutdown()
LOG.info('Shutdown request received...')
self.running = False
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
class AlertaDaemon(Daemon):
def run(self):
self.running = True
# Create internal queue
self.queue = Queue.Queue()
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=ServerMessage(self.queue))
self.mq.subscribe()
# Start worker threads
LOG.debug('Starting %s alert handler threads...', CONF.server_threads)
for i in range(CONF.server_threads):
w = WorkerThread(self.mq, self.queue)
try:
w.start()
except Exception, e:
LOG.error('Worker thread #%s did not start: %s', i, e)
continue
LOG.info('Started alert handler thread: %s', w.getName())
while not self.shuttingdown:
try:
time.sleep(0.1)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
for i in range(CONF.server_threads):
self.queue.put(None)
LOG.info('Shutdown request received...')
self.running = False
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
class LoggerDaemon(Daemon):
"""
Index alerts in ElasticSearch using Logstash format so that logstash GUI and/or Kibana can be used as front-ends
"""
logger_opts = {
'es_host': 'localhost',
'es_port': 9200,
'es_index':
'alerta-%Y.%m.%d', # NB. Kibana config must match this index
}
def __init__(self, prog, **kwargs):
config.register_opts(LoggerDaemon.logger_opts)
Daemon.__init__(self, prog, kwargs)
def run(self):
self.running = True
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=LoggerMessage(self.mq))
self.mq.subscribe(destination=CONF.outbound_queue)
while not self.shuttingdown:
try:
LOG.debug('Waiting for log messages...')
time.sleep(CONF.loop_every)
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
class LoggerDaemon(Daemon):
"""
Index alerts in ElasticSearch using Logstash format so that logstash GUI and/or Kibana can be used as front-ends
"""
logger_opts = {
'es_host': 'localhost',
'es_port': 9200,
'es_index': 'alerta-%Y.%m.%d', # NB. Kibana config must match this index
}
def __init__(self, prog, **kwargs):
config.register_opts(LoggerDaemon.logger_opts)
Daemon.__init__(self, prog, kwargs)
def run(self):
self.running = True
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=LoggerMessage(self.mq))
self.mq.subscribe(destination=CONF.outbound_queue)
while not self.shuttingdown:
try:
LOG.debug('Waiting for log messages...')
time.sleep(CONF.loop_every)
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
class PagerDutyDaemon(Daemon):
pagerduty_opts = {
'pagerduty_endpoint': 'https://events.pagerduty.com/generic/2010-04-15/create_event.json',
'pagerduty_api_key': '',
}
def __init__(self, prog, **kwargs):
config.register_opts(PagerDutyDaemon.pagerduty_opts)
Daemon.__init__(self, prog, kwargs)
def run(self):
self.running = True
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=PagerDutyMessage(self.mq))
self.mq.subscribe(destination=CONF.outbound_topic) # TODO(nsatterl): use dedicated queue?
while not self.shuttingdown:
try:
LOG.debug('Waiting for PagerDuty messages...')
time.sleep(CONF.loop_every)
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
def run(self):
self.running = True
# Create internal queue
self.queue = Queue.Queue()
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=ServerMessage(self.queue))
self.mq.subscribe()
# Start worker threads
LOG.debug('Starting %s alert handler threads...', CONF.server_threads)
for i in range(CONF.server_threads):
w = WorkerThread(self.mq, self.queue)
try:
w.start()
except Exception, e:
LOG.error('Worker thread #%s did not start: %s', i, e)
continue
LOG.info('Started alert handler thread: %s', w.getName())
def run(self):
self.running = True
self.statsd = StatsD() # graphite metrics
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=CloudWatchMessage(self.mq))
self.dedup = DeDup(by_value=True)
LOG.info('Connecting to SQS queue %s', CONF.cloudwatch_sqs_queue)
try:
sqs = boto.sqs.connect_to_region(
CONF.cloudwatch_sqs_region,
aws_access_key_id=CONF.cloudwatch_access_key,
aws_secret_access_key=CONF.cloudwatch_secret_key
)
except boto.exception.SQSError, e:
LOG.error('SQS API call failed: %s', e)
sys.exit(1)
def run(self):
self.running = True
self.queue = Queue.Queue() # Create internal queue
self.db = Mongo() # mongo database
self.carbon = Carbon() # carbon metrics
self.statsd = StatsD() # graphite metrics
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=ServerMessage(self.mq, self.queue, self.statsd))
self.mq.subscribe()
# Start worker threads
LOG.debug('Starting %s worker threads...', CONF.server_threads)
for i in range(CONF.server_threads):
w = WorkerThread(self.mq, self.queue, self.statsd)
try:
w.start()
except Exception, e:
LOG.error('Worker thread #%s did not start: %s', i, e)
continue
LOG.info('Started worker thread: %s', w.getName())
def main(self):
if CONF.heartbeat:
msg = Heartbeat(
origin=CONF.origin,
version=__version__,
)
else:
msg = Alert(
resource=CONF.resource,
event=CONF.event,
correlate=CONF.correlate,
group=CONF.group,
value=CONF.value,
severity=CONF.severity,
environment=CONF.environment,
service=CONF.service,
text=CONF.text,
event_type='exceptionAlert', # TODO(nsatterl): make this configurable?
tags=CONF.tags,
origin=CONF.origin,
threshold_info='n/a', #TODO(nsatterl): make this configurable?
timeout=CONF.timeout,
)
if CONF.dry_run:
print msg
else:
LOG.debug('Message => %s', repr(msg))
mq = Messaging()
mq.connect()
mq.send(msg)
mq.disconnect()
return msg.get_id()
class DynectDaemon(Daemon):
dynect_opts = {
'dynect_customer': '',
'dynect_username': '',
'dynect_password': '',
}
def __init__(self, prog, **kwargs):
config.register_opts(DynectDaemon.dynect_opts)
Daemon.__init__(self, prog, kwargs)
self.info = {}
self.last_info = {}
self.updating = False
self.dedup = DeDup(threshold=10)
def run(self):
self.running = True
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=DynectMessage(self.mq))
while not self.shuttingdown:
try:
self.queryDynect()
if self.updating:
self.alertDynect()
self.last_info = self.info
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
LOG.debug('Waiting for next check run...')
time.sleep(CONF.loop_every)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
self.running = False
def alertDynect(self):
for resource in self.info:
if resource not in self.last_info:
continue
if resource.startswith('gslb-'):
# gslb status = ok | unk | trouble | failover
text = 'GSLB status is %s.' % self.info[resource]['status']
if self.info[resource]['status'] == 'ok':
event = 'GslbOK'
severity = severity_code.NORMAL
else:
event = 'GslbNotOK'
severity = severity_code.CRITICAL
correlate = ['GslbOK', 'GslbNotOK']
elif resource.startswith('pool-'):
# pool status = up | unk | down
# pool serve_mode = obey | always | remove | no
# pool weight (1-15)
if 'down' in self.info[resource]['status']:
event = 'PoolDown'
severity = severity_code.MAJOR
text = 'Pool is down'
elif 'obey' not in self.info[resource]['status']:
event = 'PoolServe'
severity = severity_code.MAJOR
text = 'Pool with an incorrect serve mode'
elif self.check_weight(self.info[resource]['gslb'],
resource) is False:
event = 'PoolWeightError'
severity = severity_code.MINOR
text = 'Pool with an incorrect weight'
else:
event = 'PoolUp'
severity = severity_code.NORMAL
text = 'Pool status is normal'
correlate = [
'PoolUp', 'PoolDown', 'PoolServe', 'PoolWeightError'
]
else:
LOG.warning('Unknown resource type: %s', resource)
continue
# Defaults
group = 'GSLB'
value = self.info[resource]['status']
environment = ['PROD']
service = ['Network']
tags = dict()
timeout = None
threshold_info = None
summary = None
raw_data = self.info[resource]['rawData']
dynectAlert = Alert(
resource=resource,
event=event,
correlate=correlate,
group=group,
value=value,
severity=severity,
environment=environment,
service=service,
text=text,
event_type='serviceAlert',
tags=tags,
timeout=timeout,
threshold_info=threshold_info,
summary=summary,
raw_data=raw_data,
)
suppress = dynectAlert.transform_alert()
if suppress:
LOG.info('Suppressing %s alert', dynectAlert.event)
LOG.debug('%s', dynectAlert)
continue
if self.dedup.is_send(dynectAlert):
self.mq.send(dynectAlert)
def check_weight(self, parent, resource):
weight = self.info[resource]['status'].split(':')[2]
for pool in [
resource for resource in self.info
if resource.startswith('pool')
and self.info[resource]['gslb'] == parent
]:
if self.info[pool]['status'].split(':')[1] == 'no':
LOG.warning('Skipping %s because not serving for pool %s',
pool, self.info[pool]['status'])
continue
LOG.debug('pool %s weight %s <=> %s', pool,
self.info[pool]['status'].split(':')[2], weight)
if self.info[pool]['status'].split(':')[2] != weight:
return False
return True
def queryDynect(self):
LOG.info('Query DynECT to get the state of GSLBs')
try:
rest_iface = DynectRest()
if CONF.debug and CONF.use_stderr:
rest_iface.verbose = True
# login
credentials = {
'customer_name': CONF.dynect_customer,
'user_name': CONF.dynect_username,
'password': CONF.dynect_password,
}
LOG.debug('credentials = %s', credentials)
response = rest_iface.execute('/Session/', 'POST', credentials)
if response['status'] != 'success':
LOG.error('Failed to create API session: %s',
response['msgs'][0]['INFO'])
self.updating = False
return
# Discover all the Zones in DynECT
response = rest_iface.execute('/Zone/', 'GET')
LOG.debug('/Zone/ => %s', json.dumps(response, indent=4))
zone_resources = response['data']
# Discover all the LoadBalancers
for resource in zone_resources:
zone = resource.split('/')[
3] # eg. /REST/Zone/guardiannews.com/
response = rest_iface.execute('/LoadBalance/' + zone + '/',
'GET')
LOG.debug('/LoadBalance/%s/ => %s', zone,
json.dumps(response, indent=4))
gslb = response['data']
# Discover LoadBalancer pool information.
for lb in gslb:
fqdn = lb.split(
'/'
)[4] # eg. /REST/LoadBalance/guardiannews.com/id.guardiannews.com/
response = rest_iface.execute(
'/LoadBalance/' + zone + '/' + fqdn + '/', 'GET')
LOG.debug('/LoadBalance/%s/%s/ => %s', zone, fqdn,
json.dumps(response, indent=4))
status = response['data']['status']
monitor = response['data']['monitor']
self.info['gslb-' + fqdn] = {
'status': status,
'gslb': fqdn,
'rawData': monitor
}
for pool in response['data']['pool']:
name = '%s-%s' % (fqdn, pool['label'].replace(
' ', '-'))
status = '%s:%s:%s' % (
pool['status'], pool['serve_mode'], pool['weight'])
self.info['pool-' + name] = {
'status': status,
'gslb': fqdn,
'rawData': pool
}
LOG.info('Finished object discovery query.')
LOG.debug('GSLBs and Pools: %s', json.dumps(self.info, indent=4))
# logout
rest_iface.execute('/Session/', 'DELETE')
except Exception, e:
LOG.error('Failed to discover GSLBs: %s', e)
self.updating = False
self.updating = True
class DynectDaemon(Daemon):
dynect_opts = {
'dynect_customer': '',
'dynect_username': '',
'dynect_password': '',
}
def __init__(self, prog, **kwargs):
config.register_opts(DynectDaemon.dynect_opts)
Daemon.__init__(self, prog, kwargs)
self.info = {}
self.last_info = {}
self.updating = False
self.dedup = DeDup(threshold=10)
def run(self):
self.running = True
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=DynectMessage(self.mq))
while not self.shuttingdown:
try:
self.queryDynect()
if self.updating:
self.alertDynect()
self.last_info = self.info
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
LOG.debug('Waiting for next check run...')
time.sleep(CONF.loop_every)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
self.running = False
def alertDynect(self):
for resource in self.info:
if resource not in self.last_info:
continue
if resource.startswith('gslb-'):
# gslb status = ok | unk | trouble | failover
text = 'GSLB status is %s.' % self.info[resource]['status']
if self.info[resource]['status'] == 'ok':
event = 'GslbOK'
severity = severity_code.NORMAL
else:
event = 'GslbNotOK'
severity = severity_code.CRITICAL
correlate = ['GslbOK', 'GslbNotOK']
elif resource.startswith('pool-'):
# pool status = up | unk | down
# pool serve_mode = obey | always | remove | no
# pool weight (1-15)
if 'down' in self.info[resource]['status']:
event = 'PoolDown'
severity = severity_code.MAJOR
text = 'Pool is down'
elif 'obey' not in self.info[resource]['status']:
event = 'PoolServe'
severity = severity_code.MAJOR
text = 'Pool with an incorrect serve mode'
elif self.check_weight(self.info[resource]['gslb'], resource) is False:
event = 'PoolWeightError'
severity = severity_code.MINOR
text = 'Pool with an incorrect weight'
else:
event = 'PoolUp'
severity = severity_code.NORMAL
text = 'Pool status is normal'
correlate = ['PoolUp', 'PoolDown', 'PoolServe', 'PoolWeightError']
else:
LOG.warning('Unknown resource type: %s', resource)
continue
# Defaults
group = 'GSLB'
value = self.info[resource]['status']
environment = ['PROD']
service = ['Network']
tags = list()
timeout = None
threshold_info = None
summary = None
raw_data = self.info[resource]['rawData']
dynectAlert = Alert(
resource=resource,
event=event,
correlate=correlate,
group=group,
value=value,
severity=severity,
environment=environment,
service=service,
text=text,
event_type='serviceAlert',
tags=tags,
timeout=timeout,
threshold_info=threshold_info,
summary=summary,
raw_data=raw_data,
)
suppress = dynectAlert.transform_alert()
if suppress:
LOG.info('Suppressing %s alert', dynectAlert.event)
LOG.debug('%s', dynectAlert)
continue
if self.dedup.is_send(dynectAlert):
self.mq.send(dynectAlert)
def check_weight(self, parent, resource):
weight = self.info[resource]['status'].split(':')[2]
for pool in [resource for resource in self.info if resource.startswith('pool') and self.info[resource]['gslb'] == parent]:
if self.info[pool]['status'].split(':')[1] == 'no':
LOG.warning('Skipping %s because not serving for pool %s', pool, self.info[pool]['status'])
continue
LOG.debug('pool %s weight %s <=> %s', pool, self.info[pool]['status'].split(':')[2], weight)
if self.info[pool]['status'].split(':')[2] != weight:
return False
return True
def queryDynect(self):
LOG.info('Query DynECT to get the state of GSLBs')
try:
rest_iface = DynectRest()
if CONF.debug and CONF.use_stderr:
rest_iface.verbose = True
# login
credentials = {
'customer_name': CONF.dynect_customer,
'user_name': CONF.dynect_username,
'password': CONF.dynect_password,
}
LOG.debug('credentials = %s', credentials)
response = rest_iface.execute('/Session/', 'POST', credentials)
if response['status'] != 'success':
LOG.error('Failed to create API session: %s', response['msgs'][0]['INFO'])
self.updating = False
return
# Discover all the Zones in DynECT
response = rest_iface.execute('/Zone/', 'GET')
LOG.debug('/Zone/ => %s', json.dumps(response, indent=4))
zone_resources = response['data']
# Discover all the LoadBalancers
for resource in zone_resources:
zone = resource.split('/')[3] # eg. /REST/Zone/guardiannews.com/
response = rest_iface.execute('/LoadBalance/' + zone + '/', 'GET')
LOG.debug('/LoadBalance/%s/ => %s', zone, json.dumps(response, indent=4))
gslb = response['data']
# Discover LoadBalancer pool information.
for lb in gslb:
fqdn = lb.split('/')[4] # eg. /REST/LoadBalance/guardiannews.com/id.guardiannews.com/
response = rest_iface.execute('/LoadBalance/' + zone + '/' + fqdn + '/', 'GET')
LOG.debug('/LoadBalance/%s/%s/ => %s', zone, fqdn, json.dumps(response, indent=4))
status = response['data']['status']
monitor = response['data']['monitor']
self.info['gslb-' + fqdn] = {'status': status, 'gslb': fqdn, 'rawData': monitor}
for pool in response['data']['pool']:
name = '%s-%s' % (fqdn, pool['label'].replace(' ', '-'))
status = '%s:%s:%s' % (pool['status'], pool['serve_mode'], pool['weight'])
self.info['pool-' + name] = {'status': status, 'gslb': fqdn, 'rawData': pool}
LOG.info('Finished object discovery query.')
LOG.debug('GSLBs and Pools: %s', json.dumps(self.info, indent=4))
# logout
rest_iface.execute('/Session/', 'DELETE')
except Exception, e:
LOG.error('Failed to discover GSLBs: %s', e)
self.updating = False
self.updating = True
class SolarWindsDaemon(Daemon):
solarwinds_opts = {
'solarwinds_host': 'localhost',
'solarwinds_username': '******',
'solarwinds_password': '',
'solarwinds_group': 'websys',
}
def __init__(self, prog, **kwargs):
config.register_opts(SolarWindsDaemon.solarwinds_opts)
Daemon.__init__(self, prog, kwargs)
def run(self):
self.running = True
while True:
try:
swis = SwisClient(username=CONF.solarwinds_username,
password=CONF.solarwinds_password)
except Exception, e:
LOG.error('SolarWinds SWIS Client error: %s', e)
time.sleep(30)
else:
break
LOG.info('Polling for SolarWinds events on %s' % CONF.solarwinds_host)
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=SolarWindsMessage(self.mq))
self.dedup = DeDup(by_value=True)
while not self.shuttingdown:
try:
LOG.debug('Polling SolarWinds...')
send_heartbeat = True
# network, interface and volume events
try:
events = swis.get_npm_events()
except IOError:
events = []
send_heartbeat = False
solarwindsAlerts = self.parse_events(events)
for solarwindsAlert in solarwindsAlerts:
if self.dedup.is_send(solarwindsAlert):
self.mq.send(solarwindsAlert)
# Cisco UCS events
try:
events = swis.get_ucs_events()
except IOError:
events = []
send_heartbeat = False
solarwindsAlerts = self.parse_events(events)
for solarwindsAlert in solarwindsAlerts:
if self.dedup.is_send(solarwindsAlert):
self.mq.send(solarwindsAlert)
if send_heartbeat:
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
else:
LOG.error('SolarWinds failure. Skipping heartbeat.')
time.sleep(CONF.loop_every)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
class AlertaDaemon(Daemon):
alerta_opts = {
'forward_duplicate': 'no',
}
def __init__(self, prog, **kwargs):
config.register_opts(AlertaDaemon.alerta_opts)
Daemon.__init__(self, prog, kwargs)
def run(self):
self.running = True
self.queue = Queue.Queue() # Create internal queue
self.db = Mongo() # mongo database
self.carbon = Carbon() # carbon metrics
self.statsd = StatsD() # graphite metrics
# Connect to message queue
self.mq = Messaging()
self.mq.connect(
callback=ServerMessage(self.mq, self.queue, self.statsd))
self.mq.subscribe()
# Start worker threads
LOG.debug('Starting %s worker threads...', CONF.server_threads)
for i in range(CONF.server_threads):
w = WorkerThread(self.mq, self.queue, self.statsd)
try:
w.start()
except Exception, e:
LOG.error('Worker thread #%s did not start: %s', i, e)
continue
LOG.info('Started worker thread: %s', w.getName())
while not self.shuttingdown:
try:
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version, timeout=CONF.loop_every)
self.mq.send(heartbeat)
time.sleep(CONF.loop_every)
LOG.info('Alert processing queue length is %d',
self.queue.qsize())
self.carbon.metric_send('alerta.alerts.queueLength',
self.queue.qsize())
self.db.update_queue_metric(self.queue.qsize())
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
for i in range(CONF.server_threads):
self.queue.put(None)
w.join()
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
class PingerDaemon(Daemon):
pinger_opts = {
"ping_file": "/etc/alerta/alert-pinger.targets",
"ping_max_timeout": 15, # seconds
"ping_max_retries": 2,
"ping_slow_warning": 5, # ms
"ping_slow_critical": 10, # ms
"server_threads": 20,
}
def __init__(self, prog, **kwargs):
config.register_opts(PingerDaemon.pinger_opts)
Daemon.__init__(self, prog, kwargs)
def run(self):
self.running = True
# Create internal queue
self.queue = Queue.Queue()
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=PingerMessage(self.mq))
self.dedup = DeDup()
self.carbon = Carbon() # graphite metrics
# Initialiase ping targets
ping_list = init_targets()
# Start worker threads
LOG.debug("Starting %s worker threads...", CONF.server_threads)
for i in range(CONF.server_threads):
w = WorkerThread(self.mq, self.queue, self.dedup, self.carbon)
try:
w.start()
except Exception, e:
LOG.error("Worker thread #%s did not start: %s", i, e)
continue
LOG.info("Started worker thread: %s", w.getName())
while not self.shuttingdown:
try:
for p in ping_list:
if "targets" in p and p["targets"]:
for target in p["targets"]:
environment = p["environment"]
service = p["service"]
retries = p.get("retries", CONF.ping_max_retries)
self.queue.put((environment, service, target, retries, time.time()))
LOG.debug("Send heartbeat...")
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
time.sleep(CONF.loop_every)
LOG.info("Ping queue length is %d", self.queue.qsize())
self.carbon.metric_send("alert.pinger.queueLength", self.queue.qsize())
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info("Shutdown request received...")
self.running = False
for i in range(CONF.server_threads):
self.queue.put(None)
w.join()
LOG.info("Disconnecting from message broker...")
self.mq.disconnect()
class UrlmonDaemon(Daemon):
urlmon_opts = {
'urlmon_file': '/etc/alerta/alert-urlmon.targets',
'urlmon_max_timeout': 15, # seconds
'urlmon_slow_warning': 2000, # ms
'urlmon_slow_critical': 5000, # ms
}
def __init__(self, prog, **kwargs):
config.register_opts(UrlmonDaemon.urlmon_opts)
Daemon.__init__(self, prog, kwargs)
def run(self):
self.running = True
# Create internal queue
self.queue = Queue.Queue()
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=UrlmonMessage(self.mq))
self.dedup = DeDup()
self.carbon = Carbon() # graphite metrics
# Initialiase alert rules
urls = init_urls()
# Start worker threads
LOG.debug('Starting %s worker threads...', CONF.server_threads)
for i in range(CONF.server_threads):
w = WorkerThread(self.mq, self.queue, self.dedup, self.carbon)
try:
w.start()
except Exception, e:
LOG.error('Worker thread #%s did not start: %s', i, e)
continue
LOG.info('Started worker thread: %s', w.getName())
while not self.shuttingdown:
try:
for url in urls:
self.queue.put((url, time.time()))
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
time.sleep(CONF.loop_every)
LOG.info('URL check queue length is %d', self.queue.qsize())
self.carbon.metric_send('alert.urlmon.queueLength',
self.queue.qsize())
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
for i in range(CONF.server_threads):
self.queue.put(None)
w.join()
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
class PingerDaemon(Daemon):
pinger_opts = {
'ping_file': '/etc/alerta/alert-pinger.targets',
'ping_max_timeout': 15, # seconds
'ping_max_retries': 2,
'ping_slow_warning': 5, # ms
'ping_slow_critical': 10, # ms
'server_threads': 20,
}
def __init__(self, prog, **kwargs):
config.register_opts(PingerDaemon.pinger_opts)
Daemon.__init__(self, prog, kwargs)
def run(self):
self.running = True
# Create internal queue
self.queue = Queue.Queue()
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=PingerMessage(self.mq))
self.dedup = DeDup()
self.carbon = Carbon() # graphite metrics
# Initialiase ping targets
ping_list = init_targets()
# Start worker threads
LOG.debug('Starting %s worker threads...', CONF.server_threads)
for i in range(CONF.server_threads):
w = WorkerThread(self.mq, self.queue, self.dedup, self.carbon)
try:
w.start()
except Exception, e:
LOG.error('Worker thread #%s did not start: %s', i, e)
continue
LOG.info('Started worker thread: %s', w.getName())
while not self.shuttingdown:
try:
for p in ping_list:
if 'targets' in p and p['targets']:
for target in p['targets']:
environment = p['environment']
service = p['service']
retries = p.get('retries', CONF.ping_max_retries)
self.queue.put(
(environment, service, target, retries,
time.time()))
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
time.sleep(CONF.loop_every)
LOG.info('Ping queue length is %d', self.queue.qsize())
self.carbon.metric_send('alert.pinger.queueLength',
self.queue.qsize())
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
for i in range(CONF.server_threads):
self.queue.put(None)
w.join()
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
class GangliaDaemon(Daemon):
def __init__(self, prog):
Daemon.__init__(self, prog)
self.dedup = DeDup(by_value=True)
def run(self):
self.running = True
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=GangliaMessage(self.mq))
while not self.shuttingdown:
try:
rules = init_rules() # re-read rule config each time
self.metric_check(rules)
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
LOG.debug('Waiting for next check run...')
time.sleep(CONF.loop_every)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
def metric_check(self, rules):
for rule in rules:
# Check rule is valid
if len(rule['thresholdInfo']) != len(rule['text']):
LOG.warning(
'Skipping invalid rule %s - MUST define alert text for each threshold.',
rule['event'])
continue
# Get list of metrics required to evaluate each rule
params = dict()
if 'filter' in rule and rule['filter'] is not None:
params[rule['filter']] = 1
for s in (' '.join(rule['text']), ' '.join(rule['thresholdInfo']),
rule['value']):
matches = re.findall('\$([a-z0-9A-Z_]+)', s)
for m in matches:
if m != 'now':
params['metric=' + m] = 1
metric_filter = '&'.join(params.keys())
LOG.debug('Metric filter = %s', metric_filter)
# Get metric data for each rule
response = GangliaDaemon.get_metrics(metric_filter)
LOG.debug('Ganglia API response: %s', response)
# Make non-metric substitutions in value, thresholdInfo and text
now = int(time.time())
rule['value'] = re.sub('\$now', str(now), rule['value'])
idx = 0
for threshold in rule['thresholdInfo']:
rule['thresholdInfo'][idx] = re.sub('\$now', str(now),
threshold)
idx += 1
idx = 0
for text in rule['text']:
rule['text'][idx] = re.sub(
'\$now',
time.strftime('%Y/%m/%d %H:%M:%S', time.localtime(now)),
text)
idx += 1
metric = dict()
for m in response:
# Make metric-based substitutions in resource eg. per instance, host or cluster
resource = re.sub('\$instance', m.get('instance', '__NA__'),
rule['resource'])
resource = re.sub('\$host', m.get('host', '__NA__'), resource)
resource = re.sub('\$cluster', m.get('cluster', '__NA__'),
resource)
if '__NA__' in resource:
LOG.debug('Metric %s doesnt match resource rule %s',
m['id'], rule['resource'])
continue
LOG.debug('Metric %s matches rule %s => %s', m['id'],
rule['resource'], resource)
# Don't generate cluster alerts from host-based metrics
if 'host' in m and not '$host' in rule['resource']:
LOG.debug(
'Skipping host-based metric for cluster-based rule')
continue
# Build up info for alert if metric value triggers threshold
if resource not in metric:
metric[resource] = dict()
if 'thresholdInfo' not in metric[resource]:
metric[resource]['thresholdInfo'] = list(
rule['thresholdInfo'])
LOG.debug('Set thresholdInfo to %s',
metric[resource]['thresholdInfo'])
if 'text' not in metric[resource]:
metric[resource]['text'] = list(rule['text'])
LOG.debug('Set text to %s', metric[resource]['text'])
if m['metric'] in rule['value']:
# Determine service and environment from rule if given
if 'environment' in rule:
metric[resource]['environment'] = [rule['environment']]
else:
metric[resource]['environment'] = [m['environment']]
LOG.debug('Set environment for alert to %s',
metric[resource]['environment'])
if 'service' in rule:
metric[resource]['service'] = [rule['service']]
else:
metric[resource]['service'] = [m['service']]
LOG.debug('Set service for alert to %s',
metric[resource]['service'])
# Use raw metric value, or sum or average if aggregated metric
if 'value' in m:
v = GangliaDaemon.quote(m['value']) # raw value
elif rule['value'].endswith('.sum'):
v = GangliaDaemon.quote(
m['sum']) # aggregated sum value if "<metric>.sum"
else:
try:
v = "%.1f" % (float(m['sum']) / float(m['num'])
) # average of aggregate value
except ZeroDivisionError:
v = 0.0
LOG.debug('Value for %s on %s is %s', m['id'], resource, v)
# If no value assign rule value
if 'value' not in metric[resource]:
metric[resource]['value'] = rule['value']
metric[resource]['value'] = re.sub(
'\$%s(\.sum)?' % m['metric'], str(v),
metric[resource]['value'])
metric[resource]['units'] = m['units']
# Assign tags
metric[resource]['tags'] = list()
metric[resource]['tags'].extend(rule['tags'])
metric[resource]['tags'].append('cluster:%s' %
m['cluster'])
if 'tags' in m and m['tags'] is not None:
metric[resource]['tags'].extend(m['tags'])
# Assign graph URL
if 'graphUrl' not in metric[resource]:
metric[resource]['graphUrls'] = list()
if 'graphUrl' in m:
metric[resource]['graphUrls'].append(m['graphUrl'])
for g in rule['graphs']:
if '$host' in rule['resource'] and 'graphUrl' in m:
metric[resource]['graphUrls'].append(
'/'.join(m['graphUrl'].rsplit('/', 2)[0:2]) +
'/graph.php?c=%s&h=%s&m=%s&r=1day&v=0&z=default'
% (m['cluster'], m['host'], g))
if '$cluster' in rule['resource'] and 'graphUrl' in m:
metric[resource]['graphUrls'].append(
'/'.join(m['graphUrl'].rsplit('/', 2)[0:2]) +
'/graph.php?c=%s&m=%s&r=1day&v=0&z=default' %
(m['cluster'], g))
metric[resource]['moreInfo'] = ''
if '$host' in rule['resource'] and 'graphUrl' in m:
metric[resource]['moreInfo'] = '/'.join(
m['graphUrl'].rsplit('/', 2)
[0:2]) + '/?c=%s&h=%s' % (m['cluster'], m['host'])
if '$cluster' in rule['resource'] and 'graphUrl' in m:
metric[resource]['moreInfo'] = '/'.join(
m['graphUrl'].rsplit(
'/', 2)[0:2]) + '/?c=%s' % m['cluster']
# Substitutions for threshold info
if m['metric'] in ''.join(rule['thresholdInfo']):
LOG.debug('Text to be substituted: %s',
''.join(rule['thresholdInfo']))
if 'value' in m:
v = GangliaDaemon.quote(m['value'])
elif rule['value'].endswith('.sum'):
v = GangliaDaemon.quote(m['sum'])
else:
try:
v = "%.1f" % (float(m['sum']) / float(m['num']))
except ZeroDivisionError:
v = 0.0
idx = 0
for threshold in metric[resource]['thresholdInfo']:
metric[resource]['thresholdInfo'][idx] = re.sub(
'\$%s(\.sum)?' % m['metric'], str(v), threshold)
idx += 1
# Substitutions for text
if m['metric'] in ''.join(rule['text']):
LOG.debug('Text to be substituted: %s',
''.join(rule['text']))
if 'value' in m:
v = GangliaDaemon.quote(m['value'])
elif rule['value'].endswith('.sum'):
v = GangliaDaemon.quote(m['sum'])
else:
try:
v = "%.1f" % (float(m['sum']) / float(m['num']))
except ZeroDivisionError:
v = 0.0
if m['type'] == 'timestamp' or m['units'] == 'timestamp':
v = time.strftime('%Y/%m/%d %H:%M:%S',
time.localtime(float(v)))
LOG.debug('Metric resource text %s', metric)
idx = 0
for text in metric[resource]['text']:
metric[resource]['text'][idx] = re.sub(
'\$%s(\.sum)?' % m['metric'], str(v), text)
idx += 1
LOG.debug('end of metric loop')
for resource in metric:
LOG.debug('Calculate final value for resource %s', resource)
index = 0
try:
calculated_value = eval(metric[resource]['value'])
except KeyError:
LOG.warning(
'Could not calculate %s value for %s because %s is not being reported',
rule['event'], resource, rule['value'])
continue
except (SyntaxError, NameError):
LOG.error(
'Could not calculate %s value for %s => eval(%s)',
rule['event'], resource, metric[resource]['value'])
continue
except ZeroDivisionError:
LOG.debug(
'Could not calculate %s value for %s => eval(%s) (division by zero). Setting to 0 instead.',
rule['event'], resource, metric[resource]['value'])
calculated_value = 0
except Exception:
LOG.error(
'Could not calculate %s value for %s => eval(%s) (threw unknown exception)',
rule['event'], resource, metric[resource]['value'])
continue
LOG.debug('Calculated value for resource %s => %s', resource,
calculated_value)
# Compare final value with each threshold
for ti in metric[resource]['thresholdInfo']:
severity, op, threshold = ti.split(':')
rule_eval = '%s %s %s' % (
GangliaDaemon.quote(calculated_value), op, threshold)
try:
result = eval(rule_eval)
except SyntaxError:
LOG.error(
'Could not evaluate %s threshold for %s => eval(%s)',
rule['event'], resource, rule_eval)
result = False
if result:
event = rule['event']
group = rule['group']
value = "%s%s" % (calculated_value,
GangliaDaemon.format_units(
metric[resource]['units']))
environment = metric[resource]['environment']
service = metric[resource]['service']
text = metric[resource]['text'][index]
tags = metric[resource]['tags']
threshold_info = ','.join(rule['thresholdInfo'])
more_info = metric[resource]['moreInfo']
graph_urls = metric[resource]['graphUrls']
gangliaAlert = Alert(
resource=resource,
event=event,
group=group,
value=value,
severity=severity,
environment=environment,
service=service,
text=text,
event_type='gangliaAlert',
tags=tags,
threshold_info=threshold_info,
more_info=more_info,
graph_urls=graph_urls,
raw_data=
'', # TODO(nsatterl): put raw metric values used to do calculation here
)
if self.dedup.is_send(gangliaAlert):
self.mq.send(gangliaAlert)
break # First match wins
index += 1
@staticmethod
def get_metrics(filter):
url = "http://%s:%s/ganglia/api/v1/metrics?%s" % (
CONF.ganglia_host, CONF.ganglia_port, filter)
LOG.info('Metric request %s', url)
try:
r = urllib2.urlopen(url, None, 15)
except urllib2.URLError, e:
LOG.error('Could not retrieve metric data from %s - %s', url, e)
return dict()
if r.getcode() is None:
LOG.error('Error during connection or data transfer (timeout=%d)',
15)
return dict()
response = json.loads(r.read())['response']
if response['status'] == 'error':
LOG.error('No metrics retreived - %s', response['message'])
return dict()
LOG.info('Retreived %s matching metrics in %ss', response['total'],
response['time'])
return response['metrics']
class UrlmonDaemon(Daemon):
urlmon_opts = {
'urlmon_file': '/etc/alerta/alert-urlmon.targets',
'urlmon_max_timeout': 15, # seconds
'urlmon_slow_warning': 2000, # ms
'urlmon_slow_critical': 5000, # ms
}
def __init__(self, prog, **kwargs):
config.register_opts(UrlmonDaemon.urlmon_opts)
Daemon.__init__(self, prog, kwargs)
def run(self):
self.running = True
# Create internal queue
self.queue = Queue.Queue()
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=UrlmonMessage(self.mq))
self.dedup = DeDup()
self.carbon = Carbon() # graphite metrics
# Initialiase alert rules
urls = init_urls()
# Start worker threads
LOG.debug('Starting %s worker threads...', CONF.server_threads)
for i in range(CONF.server_threads):
w = WorkerThread(self.mq, self.queue, self.dedup, self.carbon)
try:
w.start()
except Exception, e:
LOG.error('Worker thread #%s did not start: %s', i, e)
continue
LOG.info('Started worker thread: %s', w.getName())
while not self.shuttingdown:
try:
for url in urls:
self.queue.put((url, time.time()))
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
time.sleep(CONF.loop_every)
LOG.info('URL check queue length is %d', self.queue.qsize())
self.carbon.metric_send('alert.urlmon.queueLength', self.queue.qsize())
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
for i in range(CONF.server_threads):
self.queue.put(None)
w.join()
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
class PingerDaemon(Daemon):
def run(self):
self.running = True
# Create internal queue
self.queue = Queue.Queue()
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=PingerMessage(self.mq))
self.dedup = DeDup()
self.carbon = Carbon() # graphite metrics
self.statsd = StatsD() # graphite metrics
# Initialiase ping targets
ping_list = init_targets()
# Start worker threads
LOG.debug('Starting %s worker threads...', CONF.server_threads)
for i in range(CONF.server_threads):
w = WorkerThread(self.mq, self.queue, self.dedup, self.carbon, self.statsd)
try:
w.start()
except Exception, e:
LOG.error('Worker thread #%s did not start: %s', i, e)
continue
LOG.info('Started worker thread: %s', w.getName())
while not self.shuttingdown:
try:
for p in ping_list:
if 'targets' in p and p['targets']:
for target in p['targets']:
environment = p['environment']
service = p['service']
retries = p.get('retries', CONF.ping_max_retries)
self.queue.put((environment, service, target, retries))
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
LOG.info('Ping queue length is %d', self.queue.qsize())
time.sleep(CONF.loop_every)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
for i in range(CONF.server_threads):
self.queue.put(None)
w.join()
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
class GangliaDaemon(Daemon):
def __init__(self, prog):
Daemon.__init__(self, prog)
self.dedup = DeDup(by_value=True)
def run(self):
self.running = True
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=GangliaMessage(self.mq))
while not self.shuttingdown:
try:
rules = init_rules() # re-read rule config each time
self.metric_check(rules)
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
LOG.debug('Waiting for next check run...')
time.sleep(CONF.loop_every)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
def metric_check(self, rules):
for rule in rules:
# Check rule is valid
if len(rule['thresholdInfo']) != len(rule['text']):
LOG.warning('Skipping invalid rule %s - MUST define alert text for each threshold.', rule['event'])
continue
# Get list of metrics required to evaluate each rule
params = dict()
if 'filter' in rule and rule['filter'] is not None:
params[rule['filter']] = 1
for s in (' '.join(rule['text']), ' '.join(rule['thresholdInfo']), rule['value']):
matches = re.findall('\$([a-z0-9A-Z_]+)', s)
for m in matches:
if m != 'now':
params['metric=' + m] = 1
metric_filter = '&'.join(params.keys())
LOG.debug('Metric filter = %s', metric_filter)
# Get metric data for each rule
response = GangliaDaemon.get_metrics(metric_filter)
LOG.debug('Ganglia API response: %s', response)
# Make non-metric substitutions in value, thresholdInfo and text
now = int(time.time())
rule['value'] = re.sub('\$now', str(now), rule['value'])
idx = 0
for threshold in rule['thresholdInfo']:
rule['thresholdInfo'][idx] = re.sub('\$now', str(now), threshold)
idx += 1
idx = 0
for text in rule['text']:
rule['text'][idx] = re.sub('\$now', time.strftime('%Y/%m/%d %H:%M:%S', time.localtime(now)), text)
idx += 1
metric = dict()
for m in response:
# Make metric-based substitutions in resource eg. per instance, host or cluster
resource = re.sub('\$instance', m.get('instance', '__NA__'), rule['resource'])
resource = re.sub('\$host', m.get('host', '__NA__'), resource)
resource = re.sub('\$cluster', m.get('cluster', '__NA__'), resource)
if '__NA__' in resource:
LOG.debug('Metric %s doesnt match resource rule %s', m['id'], rule['resource'])
continue
LOG.debug('Metric %s matches rule %s => %s', m['id'], rule['resource'], resource)
# Don't generate cluster alerts from host-based metrics
if 'host' in m and not '$host' in rule['resource']:
LOG.debug('Skipping host-based metric for cluster-based rule')
continue
# Build up info for alert if metric value triggers threshold
if resource not in metric:
metric[resource] = dict()
if 'thresholdInfo' not in metric[resource]:
metric[resource]['thresholdInfo'] = list(rule['thresholdInfo'])
LOG.debug('Set thresholdInfo to %s', metric[resource]['thresholdInfo'])
if 'text' not in metric[resource]:
metric[resource]['text'] = list(rule['text'])
LOG.debug('Set text to %s', metric[resource]['text'])
if m['metric'] in rule['value']:
# Determine service and environment from rule if given
if 'environment' in rule:
metric[resource]['environment'] = [rule['environment']]
else:
metric[resource]['environment'] = [m['environment']]
LOG.debug('Set environment for alert to %s', metric[resource]['environment'])
if 'service' in rule:
metric[resource]['service'] = [rule['service']]
else:
metric[resource]['service'] = [m['service']]
LOG.debug('Set service for alert to %s', metric[resource]['service'])
# Use raw metric value, or sum or average if aggregated metric
if 'value' in m:
v = GangliaDaemon.quote(m['value']) # raw value
elif rule['value'].endswith('.sum'):
v = GangliaDaemon.quote(m['sum']) # aggregated sum value if "<metric>.sum"
else:
try:
v = "%.1f" % (float(m['sum']) / float(m['num'])) # average of aggregate value
except ZeroDivisionError:
v = 0.0
LOG.debug('Value for %s on %s is %s', m['id'], resource, v)
# If no value assign rule value
if 'value' not in metric[resource]:
metric[resource]['value'] = rule['value']
metric[resource]['value'] = re.sub('\$%s(\.sum)?' % m['metric'], str(v),
metric[resource]['value'])
metric[resource]['units'] = m['units']
# Assign tags
metric[resource]['tags'] = list()
metric[resource]['tags'].extend(rule['tags'])
metric[resource]['tags'].append('cluster:%s' % m['cluster'])
if 'tags' in m and m['tags'] is not None:
metric[resource]['tags'].extend(m['tags'])
# Assign graph URL
if 'graphUrl' not in metric[resource]:
metric[resource]['graphUrls'] = list()
if 'graphUrl' in m:
metric[resource]['graphUrls'].append(m['graphUrl'])
for g in rule['graphs']:
if '$host' in rule['resource'] and 'graphUrl' in m:
metric[resource]['graphUrls'].append('/'.join(m['graphUrl'].rsplit('/', 2)[0:2])
+ '/graph.php?c=%s&h=%s&m=%s&r=1day&v=0&z=default'
% (m['cluster'], m['host'], g))
if '$cluster' in rule['resource'] and 'graphUrl' in m:
metric[resource]['graphUrls'].append('/'.join(m['graphUrl'].rsplit('/', 2)[0:2])
+ '/graph.php?c=%s&m=%s&r=1day&v=0&z=default'
% (m['cluster'], g))
metric[resource]['moreInfo'] = ''
if '$host' in rule['resource'] and 'graphUrl' in m:
metric[resource]['moreInfo'] = '/'.join(
m['graphUrl'].rsplit('/', 2)[0:2]) + '/?c=%s&h=%s' % (m['cluster'], m['host'])
if '$cluster' in rule['resource'] and 'graphUrl' in m:
metric[resource]['moreInfo'] = '/'.join(m['graphUrl'].rsplit('/', 2)[0:2]) + '/?c=%s' % m['cluster']
# Substitutions for threshold info
if m['metric'] in ''.join(rule['thresholdInfo']):
LOG.debug('Text to be substituted: %s', ''.join(rule['thresholdInfo']))
if 'value' in m:
v = GangliaDaemon.quote(m['value'])
elif rule['value'].endswith('.sum'):
v = GangliaDaemon.quote(m['sum'])
else:
try:
v = "%.1f" % (float(m['sum']) / float(m['num']))
except ZeroDivisionError:
v = 0.0
idx = 0
for threshold in metric[resource]['thresholdInfo']:
metric[resource]['thresholdInfo'][idx] = re.sub('\$%s(\.sum)?' % m['metric'], str(v),
threshold)
idx += 1
# Substitutions for text
if m['metric'] in ''.join(rule['text']):
LOG.debug('Text to be substituted: %s', ''.join(rule['text']))
if 'value' in m:
v = GangliaDaemon.quote(m['value'])
elif rule['value'].endswith('.sum'):
v = GangliaDaemon.quote(m['sum'])
else:
try:
v = "%.1f" % (float(m['sum']) / float(m['num']))
except ZeroDivisionError:
v = 0.0
if m['type'] == 'timestamp' or m['units'] == 'timestamp':
v = time.strftime('%Y/%m/%d %H:%M:%S', time.localtime(float(v)))
LOG.debug('Metric resource text %s', metric)
idx = 0
for text in metric[resource]['text']:
metric[resource]['text'][idx] = re.sub('\$%s(\.sum)?' % m['metric'], str(v), text)
idx += 1
LOG.debug('end of metric loop')
for resource in metric:
LOG.debug('Calculate final value for resource %s', resource)
index = 0
try:
calculated_value = eval(metric[resource]['value'])
except KeyError:
LOG.warning('Could not calculate %s value for %s because %s is not being reported',
rule['event'], resource, rule['value'])
continue
except (SyntaxError, NameError):
LOG.error('Could not calculate %s value for %s => eval(%s)', rule['event'], resource,
metric[resource]['value'])
continue
except ZeroDivisionError:
LOG.debug(
'Could not calculate %s value for %s => eval(%s) (division by zero). Setting to 0 instead.',
rule['event'], resource, metric[resource]['value'])
calculated_value = 0
except Exception:
LOG.error('Could not calculate %s value for %s => eval(%s) (threw unknown exception)',
rule['event'], resource, metric[resource]['value'])
continue
LOG.debug('Calculated value for resource %s => %s', resource, calculated_value)
# Compare final value with each threshold
for ti in metric[resource]['thresholdInfo']:
severity, op, threshold = ti.split(':')
rule_eval = '%s %s %s' % (GangliaDaemon.quote(calculated_value), op, threshold)
try:
result = eval(rule_eval)
except SyntaxError:
LOG.error('Could not evaluate %s threshold for %s => eval(%s)', rule['event'],
resource, rule_eval)
result = False
if result:
event = rule['event']
group = rule['group']
value = "%s%s" % (calculated_value, GangliaDaemon.format_units(metric[resource]['units']))
environment = metric[resource]['environment']
service = metric[resource]['service']
text = metric[resource]['text'][index]
tags = metric[resource]['tags']
threshold_info = ','.join(rule['thresholdInfo'])
more_info = metric[resource]['moreInfo']
graph_urls = metric[resource]['graphUrls']
gangliaAlert = Alert(
resource=resource,
event=event,
group=group,
value=value,
severity=severity,
environment=environment,
service=service,
text=text,
event_type='gangliaAlert',
tags=tags,
threshold_info=threshold_info,
more_info=more_info,
graph_urls=graph_urls,
raw_data='', # TODO(nsatterl): put raw metric values used to do calculation here
)
if self.dedup.is_send(gangliaAlert):
self.mq.send(gangliaAlert)
break # First match wins
index += 1
@staticmethod
def get_metrics(filter):
url = "http://%s:%s/ganglia/api/v1/metrics?%s" % (CONF.ganglia_host, CONF.ganglia_port, filter)
LOG.info('Metric request %s', url)
try:
r = urllib2.urlopen(url, None, 15)
except urllib2.URLError, e:
LOG.error('Could not retrieve metric data from %s - %s', url, e)
return dict()
if r.getcode() is None:
LOG.error('Error during connection or data transfer (timeout=%d)', 15)
return dict()
response = json.loads(r.read())['response']
if response['status'] == 'error':
LOG.error('No metrics retreived - %s', response['message'])
return dict()
LOG.info('Retreived %s matching metrics in %ss', response['total'], response['time'])
return response['metrics']
import sys
from flask import Flask
from alerta.common import config
from alerta.common import log as logging
from alerta.common.mq import Messaging
from alerta.server.database import Mongo
Version = '2.1.0'
LOG = logging.getLogger(__name__)
CONF = config.CONF
config.parse_args(version=Version)
logging.setup('alerta')
app = Flask(__name__)
app.config.from_object(__name__)
db = Mongo()
mq = Messaging()
mq.connect()
import views
import management.views
class IrcbotDaemon(Daemon):
ircbot_opts = {
'irc_host': 'localhost',
'irc_port': 6667,
'irc_channel': '#alerts',
'irc_user': '******',
}
def __init__(self, prog, **kwargs):
config.register_opts(IrcbotDaemon.ircbot_opts)
Daemon.__init__(self, prog, kwargs)
def run(self):
self.running = True
# An IRC client may send 1 message every 2 seconds
# See section 5.8 in http://datatracker.ietf.org/doc/rfc2813/
tokens = LeakyBucket(tokens=20, rate=2)
tokens.start()
# Connect to IRC server
try:
irc = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
irc.connect((CONF.irc_host, CONF.irc_port))
time.sleep(1)
irc.send('NICK %s\r\n' % CONF.irc_user)
time.sleep(1)
irc.send('USER %s 8 * : %s\r\n' % (CONF.irc_user, CONF.irc_user))
LOG.debug('USER -> %s', irc.recv(4096))
time.sleep(1)
irc.send('JOIN %s\r\n' % CONF.irc_channel)
LOG.debug('JOIN -> %s', irc.recv(4096))
except Exception, e:
LOG.error('IRC connection error: %s', e)
sys.exit(1)
LOG.info('Joined IRC channel %s on %s as USER %s', CONF.irc_channel, CONF.irc_host, CONF.irc_user)
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=IrcbotMessage(self.mq, irc, tokens))
self.mq.subscribe(destination=CONF.outbound_topic)
while not self.shuttingdown:
try:
LOG.debug('Waiting for IRC messages...')
ip, op, rdy = select.select([irc], [], [], CONF.loop_every)
if ip:
for i in ip:
if i == irc:
data = irc.recv(4096).rstrip('\r\n')
if len(data) > 0:
if 'ERROR' in data:
LOG.error('%s. Exiting...', data)
sys.exit(1)
else:
LOG.debug('%s', data)
else:
LOG.warning('IRC server sent no data')
if 'PING' in data:
LOG.info('IRC PING received -> PONG ' + data.split()[1])
irc.send('PONG ' + data.split()[1] + '\r\n')
elif 'ack' in data.lower():
LOG.info('Request to ACK %s by %s', data.split()[4], data.split()[0])
ack_alert(data.split()[4])
elif 'delete' in data.lower():
LOG.info('Request to DELETE %s by %s', data.split()[4], data.split()[0])
delete_alert(data.split()[4])
elif data.find('!alerta quit') != -1:
irc.send('QUIT\r\n')
else:
LOG.warning('IRC: %s', data)
else:
i.recv()
else:
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
tokens.shutdown()
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
import sys
from flask import Flask
from alerta.common import config
from alerta.common import log as logging
from alerta.common.mq import Messaging
from alerta.server.database import Mongo
Version = '2.0.1'
LOG = logging.getLogger(__name__)
CONF = config.CONF
config.parse_args(sys.argv[1:], version=Version)
logging.setup('alerta')
app = Flask(__name__)
app.config.from_object(__name__)
db = Mongo()
mq = Messaging()
mq.connect()
import views
import management.views
class AwsDaemon(Daemon):
aws_opts = {
'fog_file': '/etc/fog/alerta.conf',
'ec2_regions': ['eu-west-1', 'us-east-1'],
'http_proxy': None,
'https_proxy': None,
}
def __init__(self, prog, **kwargs):
config.register_opts(AwsDaemon.aws_opts)
Daemon.__init__(self, prog, kwargs)
self.info = {}
self.last = {}
self.lookup = {}
self.dedup = DeDup()
def run(self):
self.running = True
# Read in FOG config file
try:
self.fog = yaml.load(open(CONF.fog_file).read())
except IOError, e:
LOG.error('Could not read AWS credentials file %s: %s',
CONF.fog_file, e)
sys.exit(1)
if not self.fog:
LOG.error('No AWS credentials found in FOG file %s. Exiting...',
CONF.fog_file)
sys.exit(1)
# Connect to message queue
self.mq = Messaging()
self.mq.connect(callback=AwsMessage(self.mq))
if CONF.http_proxy:
os.environ['http_proxy'] = CONF.http_proxy
if CONF.https_proxy:
os.environ['https_proxy'] = CONF.https_proxy
while not self.shuttingdown:
try:
self.ec2_status_check()
LOG.debug('Send heartbeat...')
heartbeat = Heartbeat(version=Version)
self.mq.send(heartbeat)
LOG.debug('Waiting for next check run...')
time.sleep(CONF.loop_every)
except (KeyboardInterrupt, SystemExit):
self.shuttingdown = True
LOG.info('Shutdown request received...')
self.running = False
LOG.info('Disconnecting from message broker...')
self.mq.disconnect()
