This is a Feature Proposal

Description

If a user attempts to put a user-defined top-level property in serverless.yml (other than one called custom), it is silently ignored while packaging and anything referencing that property will fail down the line, potentially not until deployment time.

I think this issue is compounded by the fact that the documentation isn’t entirely clear on the fact that a custom top-level property in the yml file must be called custom.

• What went wrong? serverless package does not call out that an improperly named top-level property is defined and will be ignored. Instead, there is only a warning that properties referencing this top-level property have no value.

• What did you expect should have happened?

  • If custom is the only name allowed for user-defined top level properties, I would expect a warning or error at the serverless package step to alert the user that what they’re trying to do isn’t going to work.
  • I’d also expect that the documentation would have a sentence explicitly calling out that this naming restriction exists.

• What was the config you used?

Example serverless.yml exerpt:

service: foo

# This will work
custom: someValue

# This will not
bar: someValue

provider:
...

Feature proposals:

• serverless package prints a warning or error when it detects a top-level property which it’s going to ignore.
• Documentation is improved to be more explicit about the custom naming convention.
  • If this is desired, I can look at putting in a PR.

Additional Data

• Serverless Framework Version you’re using: 1.18