Is there an existing issue for this?
Use case description
Our project has a fairly large AppSync API, with a lot of resolvers.
We use RDS as data source and have had to create a few helper functions to cover our needs for resolver SQL queries. Due to the nature of AppSync JS resolvers, this means that the helper functions are bundled into each resolver template that they are used in which results in us having fairly large mapping templates when packaged.
Due to Serverless using the RequestMappingTemplate field in resolvers and function configurations, our templates have begun to hit the 1mb limit purely because of our resolver code.
This has resulted in us having to split resolvers into separate services in order for them to be deployed.
Proposed solution (optional)
A simple workaround for this issue would be to offer the ability to upload resolver and function configuration code to S3 and using the RequestMappingTemplateS3Location field. Which would prevent the unnecessary bloating of template files.
Is there an existing issue for this?
Use case description
Our project has a fairly large AppSync API, with a lot of resolvers.
We use RDS as data source and have had to create a few helper functions to cover our needs for resolver SQL queries. Due to the nature of AppSync JS resolvers, this means that the helper functions are bundled into each resolver template that they are used in which results in us having fairly large mapping templates when packaged.
Due to Serverless using the RequestMappingTemplate field in resolvers and function configurations, our templates have begun to hit the 1mb limit purely because of our resolver code.
This has resulted in us having to split resolvers into separate services in order for them to be deployed.
Proposed solution (optional)
A simple workaround for this issue would be to offer the ability to upload resolver and function configuration code to S3 and using the RequestMappingTemplateS3Location field. Which would prevent the unnecessary bloating of template files.