Online Store Project

Sales microservice

Customers microservice

Shipping microservice

foo@bar:~$ minikube start &&  minikube addons enable ingress

(k8s/mongodb)

foo@bar:~$ cd k8s/mongodb && kubectl apply -f . 

wait 1 minute aprox unil ready 1/1

foo@bar:~$ cd ../rabitmq/ 
foo@bar:~$ kubectl apply -f https://github.com/rabbitmq/cluster-operator/releases/latest/download/cluster-operator.yml
foo@bar:~$ kubectl apply -f rabbitmqcluster.yaml

Now we can deploy ours microservices. But before this, we need to build the images and push to dockerhub. The files I execute were docker-build-images.sh and docker-push-images.sh. We will use the docker images I have already push to dockerhub

foo@bar:~$ cd ../microservices/
foo@bar:~$ ./kubectl-apply-microservices.sh 
foo@bar:~$ kubectl get pods
foo@bar:~$ kubectl describe ingress online-store-ingress

With this command above, we deploy the microservices, then with get pods, I see when all of them is un running state.

192.168.49.2 is my minikube ip, get your minikube ip running in console:

foo@bar:~$ minikube ip 

now we test if customers is running ok. It must return an array without elements "[]"

foo@bar:~$ curl http://192.168.49.2/customers
[]

now we have all running in kubernates. Lets populate the DB and then run the bot

foo@bar:~$ cd ../.. && cd bot

populate delete DB data and upload again using API sales and customers microservices. I load 1000 customers with random wallet_usd and products to catalog store.

Attention => get your minikube IP and edit (bot/docker-compose.yml) with your IP (mine is 192.168.49.2)

foo@bar:~$ docker compose up populate

this is the docker-container that will use 200 customers and make each order with 1-5 products for them

foo@bar:~$ docker compose up run_bot

when it finish,

finally we get the summary and stats for customers. Go to the browser: http://192.168.49.2/reporting/json (I can see json data) http://192.168.49.2/reporting/csv-summary (It downloads a csv file) http://192.168.49.2/reporting/csv-stats-by-customer (it downloads a csv file)

you can run locust if you run. Then go to http://192.168.49.2:8089/ , select 200 users and only press Start swarming if you want, you can run again populate before.

foo@bar:~$ docker compose up run_locust

In /reporting/analysis-jupyter/analysis.ipynb you can see what I code to analyze the data from the different microservices before to code it as a downloadable file

SwaggerAPI

• http://192.168.49.2/customers/api/docs
• http://192.168.49.2/sales/api/docs
• http://192.168.49.2/shipping/api/docs
• http://192.168.49.2/reporting/api/docs

Technologies

DB Motor: I use MongoDB. I setup a MongoDB K8s Cluster, and for each microservice I created one database Web Server: FastAPI: This was the first time I use in a project, I wanted to learn it. Manage Async operations, so Motor MongoDB is the library to use with MongoDB as Async. For background processes, I worked with PyMongo directly. I used Jupyter notebook to analyze and put all together the reporting process. Then I put to work this in a microservice separated "reporting".