Unexpected Queue deletion

I just redeployed a stack that added VisibilityTimeout to a SQS FIFIO Queue resource.

This unexpectedly deleted the queue. The CloudFormation logs say

Requested update requires the creation of a new physical resource; hence creating one.

Does any change to a SQS Resource cause this to happen ? What about messages that might be in the Queue ? Does Serverless take care not to loose them ?

So either this is a rare event, or it’s common and no one knows what Serverless does here ?

Or does “everyone know” that Serverless will sometimes delete a SQS with no warning, and drop messages in flight on the floor ?