I just changed an attribute in my cognito pool (under serverless.yml resources), and deployed my stack. To my horror, this caused AWS’ cloudformation to REPLACE the cognito user pool! There’s a lot of threads on how crazy aws for doing things this way (eg that some attribute changes cause the whole pool to be replaced if you use cloudformation, whilst the same attributes can be changed via the console), but that’s beside the point here.
What I’m after is for serverless to fail the deploy, or at least warn and ask the user if you want to continue, if the deploy would cause ANY resource to get completely replaced. In particular Cognito user pools and DynamoDB tables spring to mind. It’s far too easy to change an attribute and miss the documentation of cloud formation and end up losing all your data
A plugin (or setting) to fail the deploy (unless maybe --force is given) would be great!