Tenki Forecast

Overview

Yet another amazing simple weather app

Quick Start

The app consist on 2 microservices, both of them containerized with Docker, and a reverse proxy, that manages the discovery and routing activities for both microservices. They were named as follow:

• Tenki forecast client: or the frontend web UI to use the app
• Tenki forecast server: backend that expose an API that fetchs the weather data on public APIs on Google Maps and DarkSky
• Traefik reverse proxy: that uses Traefik to handle the discovery, routing and exposure of the previous microservices

This documentation will help you to build and deploy them to Amazon ECS, which will run a cluster with an EC2 instance specially optimized to run docker containers.

Demo

You can see Tenki forecast client in action through your browser on its demo web UI

Also, you can reach Tenki forecast server running on its demo API entrypoint. You need to request it through POST method. For example:

$ curl -v -H 'Content-Type: application/json' -d '{"latitude":-33.4446699,"longitude":-70.6493836}' -X POST http://ec2-18-207-108-90.compute-1.amazonaws.com/api/tenki-forecast

Finally, you can view Traefik reverse proxy routing architecture and statistics on:

• Dashboard
• Requests status
• Health statistics

Prerequisites

You will need a AWS account to fully deploy this project. It is recommended to use a Free Tier account.

To follow CI best practices, you can activate your repository on travis.org and use the travis.yml provided file.

Also, you will need to install the following tools and technologies:

• gcc >= 5.4.0
• yarn == 1.12.3
• Node.js >= 8.12.0
• Express >= 4.16.3
• React >= 16.6.3
• Next.js == 7.0.2
• dark-sky == 1.1.4
• google-map-react >= 1.1.1
• Traefik == 1.7.4
• redis-cli >= 5.0.2
• redis-server >= 3.0.6
• aws-cli >= 1.16.59
• ecs-cli >= 1.12.0

Local development environment

• make dev-install to define and install dependencies for local development environment
• make dev-run-server to run server app on development mode
• make dev-run-client to run client app on development mode

Local production environment

• make install to define and install dependencies for local production environment
• make run-server to run server app on production mode
• make run-client to run client app on production mode

Running tests

• make test to run mocha tests on server and client microservices

Production building and deployment

First of all, you need to activate Amazon ECR, Amazon ECS services and Amazon ElastiCache.

Please adjust your AWS credentials and configuration on bin/aws-config.sh. You need to enter a VPC and its Subnets, Security Group and Instance Role.

AWS Security group

For the Security group, you need to expose traffic to the following ports:

• 3000 (to allow Tenki forecast client requests)
• 4040 (to allow Tenki forecast server requests)
• 80 (to allow Traefik reverse proxy requests and routing traffic)
• 8080 (to allow Traefik reverse proxy dashboard requests)

AWS Instance Role

For the Instance Role, you need to attach the following AWS policies:

• AmazonElastiCacheFullAccess
• AmazonEC2ContainerServiceFullAccess
• AmazonECSTaskExecutionRolePolicy
• AmazonEC2ContainerServiceforEC2Role

And create the following custom policy for Traefik to work:

Name:
traefik-policy

Contents (JSON):
    {
        "Version": "2012-10-17",
        "Statement": [
            {
                "Sid": "TraefikECSReadAccess",
                "Effect": "Allow",
                "Action": [
                    "ecs:ListClusters",
                    "ecs:DescribeClusters",
                    "ecs:ListTasks",
                    "ecs:DescribeTasks",
                    "ecs:DescribeContainerInstances",
                    "ecs:DescribeTaskDefinition",
                    "ec2:DescribeInstances"
                ],
                "Resource": [
                    "*"
                ]
            }
        ]
    }

Redis on Amazon ElastiCache

On Amazon ElastiCache, create a Redis cluster with the created role and then search for its Primary Endpoint and adjust the REDIS_HOST env variable on the docker-compose.yml file

Setup Amazon ECS cluster

Then you need to setup the Amazon ECS Cluster. To do that, please adjust your credentials and configuration on bin/aws-config.sh file and then run make setup-production. You need to do this only once per cluster.

Build and deploy

Finally, run make build to build Docker images for each microservice and send them to repositories on Amazon ECR. Then, run make deploy to scale up the instance on the Cluster, create a Service and run a Task Definition which runs and exposes the 3 microservices. You can find the deployment configuration and policies on docker-compose.yml and ecs-params.yml files.

To view the running microservices public IPs, you can use make status rule.

List of building and deployment utilities

Below is the full list of the utilities used to build and deploy the project:

• make setup-production to setup a custom cluster on Amazon ECS to be able to deploy the project. You need to adjust bin/aws-config.sh file
• make build to build Docker images and push them to Amazon ECR
• make deploy to deploy a Service with defined Task Definitions to an EC2 instance with Docker container support, through Amazon ECS
• make status to check Docker images history on Amazon ECR and Task Definitions status (RUNNING, STOPPED, DRAINING, etc) on Amazon ECS
• make scale-down to scale-down the current running Service and its Task Definitions
• make scale-up to scale-up the last created Service and its Task Definitions
• make remove-production to scale-down the current running Services and its Task Definitions and permanently delete the Amazon ECS cluster

Other utilities

• make commit better commits with Commitizen, using AngularJS's commit message convention (cz-conventional-changelog)