We’re using Serverless to deploy a lambda function in AWS. We’re finding deploying the whole stack via CloudFormation to be cumbersome, error prone and slow.
This blogpost from Serverless recommends using the
serverless deploy function command for faster deployments when you’re only updating the lambda function. However, the documentation page for the command warns against using this in production since it’ll cause the stack to go out of sync.
Experientially, we’re noticing that CloudFormation doesn’t detect any drift when we update the function. Is the documentation wrong? Is it AWS failing to detect drift even though we have drift? I’d like to use the deploy function command in our production pipeline but I’m trying to understand any caveats associated with it and have not been able to find any information beyond that one statement.
If you have any insights or experience with this, I would love to hear about it. Thanks!