Build frontend with serverless.yml environment variables before deploying about serverless-finch HOT 3 CLOSED

Hi @tgfischer! Thanks for asking about this, definitely sounds like a useful plugin. I remember having similar questions in the past and I think we opted to keep build functionality separate simply because there are a lot of different build systems and this particular plugin in mainly for deployment purposes. I would imagine both serverless-build-client and serverless-finch potentially being used within some other build system.

@linusmarco What are your thoughts here?

@tgfischer regardless of this particular issue, we'd love to have your help developing new features for serverless-finch! We have several requests sitting around 😁.

from serverless-finch.

Hey @tgfischer! Sorry for the slow response. @fernando-mc is right--the solution we have for this at the moment is the routingRules configuration parameter. This allows your front end code to hit a relative endpoint (e.g. api/getData) and then S3 redirects that to your specified API Gateway endpoint. This has worked well for me, but it would be nice to put our heads together and think of a better solution. The downsides of the current solution are:

1. S3 buckets reject any HTTP request that's not a GET before any redirects can happen
2. There's some minor overhead to the extra HTTP request that happens because of the 3XX redirect

I like the concept of having a solution like yours that doesn't require redirects, but I think we should be careful to make our solution platform/build system agnostic.

Thoughts?

from serverless-finch.

Hey @fernando-mc and @linusmarco. I think I'll keep working on my plugin separately, and focus on building the client. That way users can decide what solution works best for them. I'll keep serverless-finch as the recommended way to upload the static site to S3 though.

I'll also keep an eye out for features that I think I can contribute to for this project

from serverless-finch.