NOTE: This repo contains only the documentation for the private BoltsOps Pro repo code. Original file: https://github.com/boltopspro/demo-backend/blob/master/README.md The docs are publish so they are available for interested customers. For access to the source code, you must be a paying BoltOps Pro subscriber. If are interested, you can contact us at contact@boltops.com or https://www.boltops.com
This backend app that is useful for demonstration purposes. It simply provides an endpoint that returns the number 42. It is a Sinatra Ruby app.
To install the dependencies.
git clone git@github.com:boltopspro/demo-backend.git
cd demo-backend
bundle
Start app with bin/web. Example:
$ bin/web
== Sinatra (v2.0.5) has taken the stage on 8080 for development with backup from Puma
Puma starting in single mode...
* Version 3.12.0 (ruby 2.5.3-p105), codename: Llamas in Pajamas
* Min threads: 0, max threads: 16
* Environment: development
* Listening on tcp://0.0.0.0:8080
Use Ctrl-C to stop
The backend app runs on port 8080.
Test with curl:
$ curl localhost:8080
42
Let's get ready to deploy to ECS. We'll use AWS_PROFILE to switch between accounts. More info: Settings Setup.
Create the ECR repos on both AWS acccounts.
AWS_PROFILE=dev aws ecr create-repository --repository-name demo/backend
AWS_PROFILE=prd aws ecr create-repository --repository-name demo/backend
If you need to list the repo again:
AWS_PROFILE=dev aws ecr describe-repositories
AWS_PROFILE=prd aws ecr describe-repositories
If you are using a Single AWS Account, use the commands: One Account: Create ECR Repos.
Let's deploy the demo app to an ECS cluster. We'll use the ufo tool for deployment. The project has already been set up with starter .ufo files for you. We'll just need to adjust some of the .ufo settings.
Adjust the .ufo/settings.yml file, which looks like this:
development:
image: 112233445566.dkr.ecr.us-west-2.amazonaws.com/demo/backend
aws_profile: dev
production:
image: 778899001122.dkr.ecr.us-west-2.amazonaws.com/demo/backend
aws_profile: prdAdjust the .ufo/settings.yml file to use the ECR repos that were created in the previous step.
If you are using a Single AWS Account, follow the instructions in: One Account: Adjust ECR Repos.
Next, adjust the network settings that the ECS service will use.
Since the backend app is meant to be an internal application, everything should run on PrivateApp subnets. Both the ELB and the ECS tasks will be placed on PrivateApp subnets. Adjust the .ufo network related files:
Here's what the network/development.yml settings file looks like:
vpc: vpc-111
ecs_subnets: # required: at least 2 subnets
- subnet-111,subnet-222,subnet-333
elb_subnets: # defaults to same subnets as ecs_subnets when not set
- subnet-111,subnet-222,subnet-333We can grab the values from the CloudFormation Console Outputs tab of the VPC. Here's an example of the development VPC.
Repeat the process for the network/production.yml file.
Creating an internal route53 record is optional but is recommended so you have a human-friendly name to refer to the internal backend server like http://demo-backend-web.dev.private/ Refer to the AWS Creating a Private Hosted Zone docs to create a private hosted zone named private. Here's a screenshot of the Route53 console interface for the development environment:
Create the Route53 Private Hosted Zone for records the production environment also.
Adjust these .ufo CloudFormation template related files so that ufo will create and manage the route53 DNS endpoints.
There is already an commented out dns entry at the bottom. You just have to uncomment the setting. Here's the cfn/development.yml setting:
dns:
name: "{stack_name}.dev.private."
hosted_zone_name: dev.private. # dont forget the trailing period
TTL: '60' # ttl has special upcase casingThis setting tells ufo to create the demo-backend-web.dev.private route53 record and point it to the internal ELB. Repeat the process for the production account with the cfn/production.yml file.
If you are using a Single AWS Account, follow the instructions in: One Account: Internal Route53 Record.
We're ready to deploy the application with ufo ship. For convenience, we'll provide example commands for both development and production environments.
UFO_ENV=development ufo ship demo-backend-web --no-wait
UFO_ENV=production ufo ship demo-backend-web --no-wait
Check the CloudFormation and ECS console for the deployment to complete. It usually take about 5 to 6 minutes. After deployment completes you can grab the DNS endpoint from the ufo managed CloudFormation stack Outputs. Here's an screenshot example of development:
You can also use the ufo ps to get the internal ELB DNS endpoint. Examples:
$ UFO_ENV=development ufo ps demo-backend-web
Elb: internal-demo-ba-Elb-AAABBBEXAMPLE-1122334455.us-west-2.elb.amazonaws.com
Route53: demo-backend-web.dev.private
$ UFO_ENV=production ufo ps demo-backend-web
Elb: internal-demo-ba-Elb-XXXYYYEXAMPLE-1122334455.us-west-2.elb.amazonaws.com
Route53: demo-backend-web.prd.private
Since we created an internal ELB, we must be inside the VPC network to reach it. The instances that were created by the ecs-spot or ecs-asg blueprints are inside the VPC network.
Use session-manager hop onto the instance. Here's an example checking the development environment.
$ AWS_PROFILE=dev aws ssm start-session --target i-1122334455
# curl http://internal-demo-ba-Elb-AAABBBEXAMPLE-1122334455.us-west-2.elb.amazonaws.com
42
# curl http://demo-backend-web.dev.private
42
#
Repeat the process for the production account.
That's it. Go back to the main boltopspro/reference-architecture