How to set up environments for serverless functions?

Hey!

As we discussed recently here, similar to the standard development process, what you could do is to use a .env files with the required values.

You can template variables in project.yml from local .env files or App Platform environment variables.

You can substitute in variables from the runtime environment using the format "${SECRET_KEY}", where SECRET_KEY is the name of the variable specified in App Platform or your .env file:

environment: {}
parameters: {}
packages:
  - name: sample
    environment: {}
    parameters: {}
    annotations: {}
    functions:
      - name: hello
        binary: false
        main: ''
        runtime: 'python:default'
        web: true
        parameters: {}
        environment:
          SECRET_KEY: "${SECRET_KEY}"
        annotations: {}
        limits: {}

That way you never commit your .env file to GitHub and you control what the values are based on your environment.

By default, the doctl serverless deploy command will look for an .env file in the same directory as the project.yml file. You can also add an --env <path to .env-formatted file> flag to specify an alternate file name and location anywhere on the filesystem. So that way you could have .env.prod and .env.dev with different values.

Alternatively, if there are bigger differences in your dev and prod environments, you can even have separate project.yaml files. For example, you could have project.staging.yml and project.live.yml, each configured with appropriate environment variables. Then when deploying using the doctl, you can use the different files by passing the --config flag:

doctl serverless deploy package-name --config project.staging.yml

For more information about the available flags, you can take a look at the documentation here:

https://docs.digitalocean.com/reference/doctl/reference/serverless/deploy/

Let me know if you have any questions and if this works for you!

Best,

Bobby