Monitoring.md

Monitoring

Heart Beat Producer

def start_heart_beat():

    thread = threading.Thread(target=heart_beat)
    thread.start()

Topic Name

• Node

  node_ip = get('https://api.ipify.org').text
  
  topic_name = "{}-{}_{}".format("node",node_ip, "8004")

• Application

  topic_name = "{}-{}".format("application", App_instance_id)

• Service

  topic_name = "{}-{}".format("service", service_name)

heart_beat = "1"
producer.send(topic_name, heart_beat)

---

Heart Beat Consumer

threading.Thread(target=receive_reg_dereg_request).start()

Register & Deregister

• Service, Node & Application will register themselves dynamically on event.

def receive_reg_dereg_request():

	reg_dereg_topic = 'register_deregister'
	for message in consumer:
		
		if request_type == 'register':
			register(request_msg)
		else:	
			deregister(request_msg)

• Register

  • Indirectly Start Listening heart-beat coming from That node/ Application/ Service on new thread.

  def register(request_msg):
  
  	# request_msg --> Topic Name (Described in Heart Beat Producer)
  	
  	threading.Thread(
  		target=heart_beat_consumer, 
  		args=(topic_name)
  	).start()

• Deregister

  # Applicable Only to Applications
  # Remove from Database

Heart Beat Consumer

def heart_beat_consumer(topic_name):
	consumer = KafkaConsumer(
					topic_name, 
					['{}:{}'.format(kafka_ip, kafka_port)], 
					consumer_timeout_ms = 16000
				)
	
	for message in consumer:
		pass

	inform_fault_tolerance(topic_name)

def	inform_fault_tolerance(topic_name):

	service_name = "Node/ Service/ App_instance_ID"
	topic_name = "monitoring"
	
	producer.send(topic_name, service_name)