When using a shared API Gateway, there are times we want to share existing models. One option is to use Cloud formation cross stack reference. So I create the model and declare the Output in the main api gateway, and in the service I want to reference it, I use:
The problem with this approach is that when using the !ImportValue, instead of a string, this model will become an object like this:
{ 'Fn::ImportValue': 'MyExistingResponseModel' }
When adding the model dependency to Cloud Formation, the function will add it like this [object Object]Model and return an error.
Since this model already exists, so we don't need to add it to the DependsOn property. Assuming that whenever we use !ImportValue, it will generate an object, we can add a check to the addModelDependencies function to only add it when it is a string.
When using a shared API Gateway, there are times we want to share existing models. One option is to use Cloud formation cross stack reference. So I create the model and declare the Output in the main api gateway, and in the service I want to reference it, I use:
responseModels: application/json: !ImportValue MyExistingResponseModelThe problem with this approach is that when using the !ImportValue, instead of a string, this model will become an object like this:
{ 'Fn::ImportValue': 'MyExistingResponseModel' }When adding the model dependency to Cloud Formation, the function will add it like this [object Object]Model and return an error.Since this model already exists, so we don't need to add it to the DependsOn property. Assuming that whenever we use !ImportValue, it will generate an object, we can add a check to the addModelDependencies function to only add it when it is a string.