Small personal app for learning Golang, idea was taken from the Kubernetes/Docker repo, but done in a more code involved way.
Application integrates with AWS DynamoDB for adding, viewing, updating, and deleting tasks. The core aspect of the application comes from the html/template and net/http in-built packages for Go, providing dynamic HTML pages and serving a HTTP server. Mux was used for HTTP path-based routing to the varying pages and providing HTTP endpoints to respond with the appropriate JSON.
The previous iterations of this app were made for Go learning purposes, these are placed into the legacycode folder. The first instance of the application was primarily writing to a JSON file, the second was an interactive CLI that integrated with AWS.
A simple menu provides basic navigation to various pages.
Tasks are viewed in a table which is dynmically generated upon visiting the page, the data is pulled from the AWS DynamoDB table after clicking the View Tasks menu option, this calls the relevant Go function.
Terraform was used to deploy the web application onto a custom AMI, although this custom AMI was not particularly necessary once utilising user data in AWS, as Golang and Git could be installed on the machine at boot-time. The Terraform file creates a VPC for deploying the instances, with an Elastic Load Balancer that listens on port 8080, this forwards the incoming traffic to the main subnet, where the instances are serving the web application. The load balancer is the internet facing portion of the architecture; the instances are not directly accessible via HTTP, they only allow traffic from 10.0.0.0/16, which is the CIDR block allocated to the VPC. SSH is enabled from anywhere for testing purposes, although it is no longer necessary as user data provides the relevant configuration to the instances.
A simple HTTP API is also available with the application to perform adding a task using a POST request, sending JSON that contains the task name, details, and completion date, with the taskID being automatically generated for the user. All tasks are also viewable using a GET request, this returns an array of JSON blobs which can be unmarshaled into the slice of Tasks, containing a struct of a singular Task. Once deployed, these are available at the HTTP endpoints of:
/api/ViewAll/api/Add
This portion serves as my own notes on configuring HTTPS/SSL, as was done with using a registered domain for testing purposes.
- SSL certificate generated through Amazon Certificate Manager, this is then loaded onto the ELB.
- ELB security group opened to also allow 443 from source 0.0.0.0/0. Alongside keeping the standard HTTP port of 80.
- The load balancer has an extra listener on port 443, this maps to the instance port of 8080 (SSL Termination).
- DNS Record in Route53 created, an A record mapping domain to the load balancer (uses alias to point to ELB auto-generated domain name).