Schema and Resolver Pre-compile

[AccsoSG]

Is your feature request related to a problem? Please describe.
Schemas and resolvers are regenerated at every server startup. Especially with a large schema, this leads to long startup times.

For use cases like autoscalling and cold starts (serverless) this is a big problem. The startup can take several seconds for large schemas, making the use of autoscalling and cold starts impractical.

Describe the solution you'd like
As already discussed in #243 (@mathix420) the possibility of a pre-compile would be perfect to solve this problem. For example, this could be solved via a specialized Neo4j GraphQL plugin for graphql-codegen.

Describe alternatives you've considered
?

[jroith]

I totally agree, for me this is also a very big issue.

[hadyrashwan]

yes me too. neo4j services are the slowest from my side.

[mathix420]

This is an issue for me as well, but this issue is a duplicate of #243, please express your interest there by adding a reaction (👍) to the first post.

[AccsoSG]

@mathix420
Actually, I disagree a little bit. I don't think it's a duplicate. Partially the same thing is talked about in 243, but the main problem there is the fact that the generation is slow and that it happens twice. Look at the title of the ticket and the labels. We are not interested in increasing the performance of the generation, but in precompiling the schema. The actual idea is unfortunately hidden in the comments of the ticket.

I just put the discussion topic in a new ticket and marked it as a feature request.

Nevertheless, I would also like the problem to be addressed, as it could be a showstopper if it is not solved. I have therefore of course also left my thumbs up there. 👍

[mathix420]

Got it, so you might be interested by my comment there #243 (comment)