Automatización de la lectura de medidores analógicos de agua basada en AIoT para asegurar la normativa ambiental de la Universidad ICESI

Contexto

Actualmente, el personal de Planta Física de la Universidad Icesi realiza registros manuales periódicos de los medidores de agua analógicos en la propiedad universitaria. La institución alberga diez aljibes, según la información proporcionada por la oficina de Planta Física y Gestión Ambiental. De estos, siete aljibes tienen acceso a la red de internet pública de la universidad, mientras que los otros tres, ubicados en zonas alejadas, no permiten la recolección automatizada de datos debido a la falta de acceso a esta red.

Objetivos

General

Desarrollar y validar un sistema tele-informático que permita automatizar el proceso de registro de lecturas de medidores analógicos de agua de los aljibes en la Universidad Icesi, para asegurar el cumplimiento de la normativa ambiental.

Específicos

1. Validar el funcionamiento del sistema AIoT de medición implementado previamente.
2. Proponer una arquitectura que permita escalar el sistema AIoT en todo el campus de la Universidad Icesi.
3. Implementar una infraestructura de comunicación y almacenamiento de datos para la medición de los nodos de AIoT.
4. Desarrollar un módulo de reportes, alertas y descarga de datos basados en el consumo de agua de los aljibes.

Network

Infrastructure

IOT 4 Layers

AIoT Flow

C4 Model

ESP32 Cam

MQTT Broker

Event Store Broker

Web App

Resilience

Deployment

1. Install Minikube

2. Starting Minikube

   minikube start --memory 6144

3. Enable Metrics k8s api

   minikube addons enable metrics-server

4. Install ingress controller (nginx), for accesing to cluster from single point (Security).

   minikube addons enable ingress

5. Creation of namespaces

   kubectl create namespace kafka
   kubectl create namespace monitoring
   kubectl create namespace apps

6. Set up the ingress to route traffic to the following services (Mqtt broker Bridge and Web App)

   kubectl apply -f ./ingress/ingress-configmap.yaml -n ingress-nginx
   kubectl apply -f ./ingress/ingress-app-rules.yaml -n apps
   # Add the port to nginx service (Edit yaml) Node Port
      #- name: mqtt
      #   port: 1883
      #   protocol: TCP
      #   targetPort: 1883

7. Deploy mongo database rs cluster with its mongo express web ui

      # Using Percona Server for MongoDB Operator
      # kubectl apply --server-side -f https://raw.githubusercontent.com/percona/percona-server-mongodb-operator/v1.15.0/deploy/bundle.yaml --namespace apps
      # Using helm bitnami chart
      # helm upgrade --install mongodb -f ./storage/mongo-values.yaml oci://registry-1.docker.io/bitnamicharts/mongodb --namespace apps
      # kubectl apply -f ./storage/mongo-crd.yaml --namespace apps
      # helm uninstall mongodb --namespace apps
      # Using kubedb operator (https://appscode.com/issue-license/) Ask for a license txt free
      # helm install kubedb oci://ghcr.io/appscode-charts/kubedb \
      # --version v2023.12.11 \
      # --namespace apps --create-namespace \
      # --set-file global.license=./license/kubedb.txt \
      # --wait --burst-limit=10000 --debug
      # helm uninstall kubedb --namespace apps
      # kubectl create -f ./storage/mongo-operator.yaml --namespace apps
      # kubectl delete -f ./storage/mongo-operator.yaml --namespace apps
      # Using local mongo db - Move to storage folder
      ./generate.sh
      ./configure.sh admin
      kubectl apply -f ./storage/mongo-ui.yaml --namespace apps
      # Creation of cisterns db and measurements colleciton from ui

8. Create configmap for jmx metrics:

   kubectl apply -f ./message-broker/kafka-metrics-config.yaml -n kafka
   kubectl apply -f ./message-broker/zookeeper-metrics-config.yaml -n kafka

9. Deploy Strimzi Kafka Operator (including ClusterRole, ClusterRoleBinding and CRDs):

   kubectl create -f 'https://strimzi.io/install/latest?namespace=kafka' -n kafka

10. Add our custom kafka resource and wait for it to be ready:

kubectl apply -f ./message-broker/kafka.yaml -n kafka
kubectl wait kafka/my-cluster --for=condition=Ready --timeout=300s -n kafka

11. Topics Creation:

kubectl apply -f message-broker/topic.yaml -n kafka

12. Create the kafka mongo connect (optional)

kubectl apply -f ./message-broker/kafka-mongo-connect.yaml -n kafka

13. Create the kafka mongo sink (optional)

kubectl apply -f ./message-broker/kafka-mongo-sink.yaml -n kafka

14. Create mqtt bridge

kubectl apply -f ./mqtt-broker -n kafka

15. Create web tier and app tier

kubectl apply -f ./compute/client/cluster/app.yaml -n apps
kubectl apply -f ./compute/server/cluster/app.yaml -n apps

16. Test our mqtt bridge

   # pip install paho-mqtt
python3 ./testing/mqtt_test.py

Monitoring (Optional)

17. Deploy Prometheus Operator (including ClusterRole, ClusterRoleBinding and CRDs):

kubectl apply -f ./monitoring/prometheus/prometheus-operator-deployment.yaml -n monitoring --force-conflicts=true --server-side

18. Deploy Prometheus

kubectl apply -f ./monitoring/prometheus/prometheus.yaml -n monitoring

19. Deploy Pod monitor

kubectl apply -f ./monitoring/strimzi-pod-monitor.yaml -n monitoring

20. Deploy grafana Enter in grafana app port forwarding 3000 to 3000 and the add prometheus datasource with the url http://prometheus-operated:9090 and also add dashboards from the folder ./dashboard (Exporter and kafka)

kubectl apply -f ./monitoring/grafana/grafana.yaml -n monitoring

21. Kafka UI

helm repo add kafka-ui https://provectus.github.io/kafka-ui-charts
helm install kafka-ui kafka-ui/kafka-ui \
   --set envs.config.KAFKA_CLUSTERS_0_NAME=my-cluster \
   --set envs.config.KAFKA_CLUSTERS_0_BOOTSTRAPSERVERS=my-cluster-kafka-bootstrap:9092 \
   --namespace kafka

22. Port forward to grafana and prometheus

kubectl port-forward svc/grafana 3000:3000 -n monitoring
kubectl port-forward svc/prometheus-operated 9090:9090 -n monitoring

K9s

Delete resources

#App
kubectl -n apps delete -f ./compute/cluster/client/app.yaml
kubectl -n apps delete -f ./compute/cluster/server/app.yaml
#Broker
kubectl -n kafka delete -f ./mqtt-broker
kubectl -n kafka delete -f ./message-broker/topic.yaml
kubectl -n kafka delete -f ./message-broker/kafka-mongo-connect.yaml
kubectl -n kafka delete -f ./message-broker/kafka-mongo-sink.yaml
kubectl -n kafka delete -f ./message-broker/kafka.yaml 
kubectl -n kafka delete -f 'https://strimzi.io/install/latest?namespace=kafka'
#Storage
kubectl -n apps delete -f ./storage/mongo-ui.yaml
kubectl -n apps delete -f ./storage/mongo.yaml
kubectl delete -f https://raw.githubusercontent.com/percona/percona-server-mongodb-operator/v1.15.0/deploy/bundle.yaml --namespace apps
#Networking
kubectl -n apps delete -f ./ingress/ingress-app-rules.yaml
kubectl -n ingress-nginx delete -f ./ingress/ingress-configmap.yaml
#Isolation
kubectl delete namespace kafka
kubectl delete namespace monitoring
kubectl delete namespace apps
#Minikube
minikube addons disable ingress
minikube addons disable metrics-server
minikube stop
# Monitoring
helm uninstall kafka-ui -n kafka
kubectl -n monitoring delete -f ./monitoring/grafana/grafana.yaml
kubectl -n monitoring delete -f ./monitoring/strimzi-pod-monitor.yaml
kubectl -n monitoring delete -f ./monitoring/prometheus/prometheus.yaml
kubectl -n kafka delete -f ./message-broker/kafka-metrics-config.yaml
kubectl -n kafka delete -f ./message-broker/zookeeper-metrics-config.yaml
kubectl -n monitoring delete -f ./monitoring/prometheus/prometheus-operator-deployment.yaml 

Flow Evidences