graphql-client VS graphql-batch

An example of a Snowflake API request using Ruby. This example uses the Ruby standard library, so it doesn’t require additional dependencies; however, in production, you could use a library like GitHub’s own graphql-client for stronger typing.

This initializer is a straightforward client, and there are some improvements you can make here. First, it's never a good idea to hard-code API secrets in your code. A better option would be to use Rails secrets or environmental variables. Second, you can load the schema just once and dump it to disk to faster read after. There are examples in the library documentation on how to make these improvements. For our little project, we will keep it simple.

A couple of examples: graphql-client a Ruby client

After recognizing the problem, we brainstormed options to offload some of the work onto the database server to ultimately reduce the Rails application’s memory consumption. One particularly promising avenue involved SQL window functions. After deciding to pursue SQL window functions, we started our work by considering the WindowKeyLoader example described in the graphql-batch repository.

Interesting, I've just been researching Dataloader implementations for Ruby and have between trying to decide between GraphQL::Dataloader, graphql-batch and BatchLoader. I'll give this a look as well. Can you also make API calls inside the loaders?

This is mainly addressed in the last thing but we cache pretty heavily on our REST APIs and for GraphQL we use the graphql-batch gem pretty heavily. Those two things can go a long way. Make sure you're using an APM like Scout to keep an eye on things and then debug the outliers.