Cloud Formation Templates for deploying Server less Applications
The template build-pipeline.yaml can be used to deploy code from GitHub to AWS CodePipeline.
The pipeline has threee stages namely Source, Build and Deploy
- In Source, the code is fetched from GitHub upon a commit to master branch (can be switched to any other branch)
- In build, the dependencies will be built based on the instructions in buildspec.yaml in the respective repository. You may customize this step to have other functionalities too. Such as running unit tests, linter etc
- In Deploy, the code is deployed using CloudFormation. This step creates a stack in which all the resources and roles mentioned in the
template.yamlof the repository are created.
Prior to creating pipeline, you have to create a GitHub Access Token for the following field
OAuthToken: "{{resolve:ssm:/github/polimetla_access_token:1}}"
To generate this token, Go to your GitHub Settings -> Developer Settings -> Personal Access Tokens -> Generate new token
In my case, I kept this token in my SSM Parameter store since this is sensitive information but as a string parameter. Saving this as a secure string would cause an error since CloudFormation cannot fetch a secure string as of now. This is still secure especially when only the Admin has the access to view/update SSM Parameters
Uploading build-pipeline.yaml with some parameter entries is the only manual step needed to deploy a CodePipeline.
Make sure that the user who uploads the templates in Cloud Formation has access to read from SSM Parameter store
Later on, upon each commit to master (or any other if you configure it that way) branch of the repository which uses this pipeline, the deployment process is automated in a CI/CD way.
CI/CD pipeline for dockerized applications