Building multi-stage Docker images with environment variables

[davidkeaveny]

Up until now, I have been building lambda functions based on .NET 8 and deploying via zip files. My lambda function consumes NuGet packages from a private NuGet feed, and it's easy enough to configure the builds for that.

I now would like to try using Docker instead of zip for the lambda functions; I've been able to use the private NuGet feed for other Docker images, I just need to inject environment variables for the username and password to use with the feed; however, I can't see anything in the documentation for dotnet lambda package or its cousins on how to pass in environment variables to the Docker build that is being invoked. Is this possible?

If I were to build my lambda in a GitHub Action and wanted to push to ECR, what would that look like in terms of the commands that I need to execute? All the documentation I can find at the moment seems to assume you're deploying from the desktop, which is fine for demo apps, not so good for real-world applications where you have tighter security requirements...

[bhoradc]

Hi @davidkeaveny,

Thank you for submitting this Discussion. We have a Feature Request #329 in backlog to Add options for pre-commands and environment variables to lambda package command. When this gets implemented, you should be able to build lambda directly in Github Action.

Regards,
Chaitanya