Add postgraphile cache to PostGraphile handler.

[RoryKelly]

I am using GraphQL mesh in a lambda.

The postgraphile handler introspects the database on start up, leading to slow lambda startup times. Postgraphile has a workaround for this documented [here]§(https://www.graphile.org/postgraphile/deploying-lambda/)

I imagine this as a config on in the yaml

sources:
  - name: database
    handler:
      postgraphile:
        connectionString: ${DATABASE_URL}
        introspectionCache: true

Then on graphql-mesh generate-sdk it also generates a cache.

[RoryKelly]

I think a similar issue exists for the graphql handler, It makes a request on startup to load the schema. Would be nice to manually add a schema file.

I think this will be an issue for a variety of handlers when deployed to lambda. Would be great to have a Cli feature where handlers can create a cache of the data the require for startup. Will make new issue

[ardatan]

You're right! We need to implement introspection caching for other handlers, too. For now, in the next release we will have cacheIntrospection option for neo4j,postgraphile and graphql handlers. Others will have this soon!

[ardatan]

Available in https://github.com/Urigo/graphql-mesh/releases/tag/v0.2.7 !