Microservice Web Application

A front end web application that connects to 5 microservices:

• Broker - single point of entry to microservices
• Authentication - PostgreSQL database for handling user auth
• Logger - MongoDB
• Mail - sends an email with specified template
• Listener - consumes messages in RabbitMQ and initiates a process

A front end html template is used to interface between the microservices.

Tools

• Go
• Docker
• Docker Swarm
• PostgresQL
• MongoDB
• RabbitMQ
• Makefile
• HTML
• Mailhog
• gRPC
• Kubernetes

Microservices in Depth

Broker Service

Handles requests from the front end application. The Broker will process the request and send a response back to the front end application.

Authentication Service

Allows the user to authenticate via the Broker Service. Data is persisted to a PostgresQL instance within Docker. This Authentication service can be used alongside any other microservice.

Logger Service

Only available within the overall microservice Kubernetes cluster. It stores all information within a MongoDB database.

Mail Service

A mail server which allows the Broker or Listener service to send email notifications.

Listener Service

Retrieves events/requests from RabbitMQ for consumption by the other microservices.

Logging connections with gRPC

In many cases, logging speed can be sped up using gRPC over JSON. The Logger service will contain a Listener for accepting gRPC connections made from the Broker via the Frond End.

The following command generates the appropriate grpc code from the logs.proto files:

• protoc --go_out=. --go_opt=paths=source_relative --go-grpc_out=. --go-grpc_opt=paths=source_relative logs.proto

Deploying to Docker Swarm

cd into each service directory, then build and push to docker hub.

• docker build -f logger-service.dockerfile -t <docker_account>/logger-service:1.0.0 .

• docker push <docker_account>/logger-service:1.0.0

• docker build -f broker-service.dockerfile -t <docker_account>/broker-service:1.0.0 .

• docker push <docker_account>/broker-service:1.0.0

• docker build -f authentication-service.dockerfile -t <docker_account>/authentication-service:1.0.0 .

• docker push <docker_account>/authentication-service:1.0.0

• docker build -f mail-service.dockerfile -t <docker_account>/mail-service:1.0.0 .

• docker push <docker_account>/mail-service:1.0.0

• docker build -f listener-service.dockerfile -t <docker_account>/listener-service:1.0.0 .

• docker push <docker_account>/listener-service:1.0.0

• docker build -f front-end.dockerfile -t <docker_account>/front-end:1.0.0 .

• docker push <docker_account>/front-end:1.0.0

• docker build -f caddy.dockerfile -t <docker_account>/micro-caddy:1.0.0 .

• docker push <docker_account>/micro-caddy:1.0.0

Deploy to Docker Swarm: docker stack deploy -c swarm.yml myapp view services: docker service ls Scaling services example: docker service scale myapp_listener-service=3

docker swarm init docker stack deploy -c swarm.yml myapp

Deploying to Kubernetes

Start Minikube: minikube start --nodes=2

Running a dockerized Postgres instance:

• docker-compose -f postgres.yml up -d

There are a few ways to expose the cluster to the local network:

Exposing the broker-service as a Load Balancer

• kubectl expose deployment broker-service --type=LoadBalancer --port=8080 --target-port=8080
• minikube tunnel

Ingress to cluster using Nginx

• minikube addons enable ingress