An attempt to get a demo for Serverless running, plus notes an opinionated structure/process for running Serverless, with a good developer experience, deployment experience and monitoring experience. For now this assumes using AWS as the provider.
- User facing parts of the product are outside of the Serverless setup - thinking of Single Page Apps or landing pages, though they can use serverless plugins to deploy in a similar way.
- Low-latency development environment, allowing new developers to make changes quickly, existing developers to be happy that they are adding features in an environment similar to production.
- Where possible development should be done with the
devenvironment within AWS, rather than having to rely on faking resources like API Gateway or Kinesis streams etc locally. Code should still be structured in such a way to allow unit tests regardless of environment and have a universal way to mock HTTP requests, stream events etc. Developers should also be allowed to deploy to their owndevenvironments for short lived testing. - Idea of multiple
services, which take care of their own resources, own development and production environment, as well as any monitoring alarms and logs if configurable using - Ability for each service to make use of shared infrastructure (via a Terraform powered
infrastructurerepo for example) such as a Kinesis stream, shared environment variables, and have knowledge of each other's service endpoints if exposed. - Deployment to production should involve some sort of CI pipeline, and in that case you can take advantage of build scripts to inject production secrets.
- Logs should be easily accessible and alarms easily configurable.
- Create a user in AWS IAM with the correct policy for your role (e.g. developer/admin - TODO manage policies as terraform?). Update
~/.aws/credentials - Make a copy of the .envrc-example to .envrc and add your the name of the aws profile from the previous step.