implement custom logics in custom resolvers VS new separate data source

[mnlbox]

Hi,

This is my use case:
I had a custom resolver that uses mesh SDK to combine two services in a new resolver. So I can call this custom resolver findUserGroups and inside this resolver implementation, I'm using two SDK functions to findUser based on email and get the id for the user and I want to pass this id as a request header to the second API call. The second one is another SDK function called findGroup each SDK functions belongs to different sources and different GraphQL subgraphs. Something like this:

import {
  FindUserResult,
  FindGroupResult,
  Resolvers,
} from "../.mesh";

const resolvers: Resolvers = {
  Query: {
    findUserGroups: async (root, args, context, info) => {
      // @ts-ignore
      const apiKey = context.headers["api-key"];
      // @ts-ignore
      const role = context.headers["role"];

      // 1. Check the required headers
      if (!apiKey || !role) {
        throw new Error(
          "api-key and role are mandatory"
        );
      }  
      
      const findUserResult: FindUserResult =
        await context.User_API.Query.findUser({
          root,
          args,
          context,
          info
        });

        const userId = findUserResult[0].id;
        console.log(userId)        

        const findGroupResult: FindGroupResult =
        await context.Group_API.Query.findGroup({
          root,
          args,
          context,
          info
        });
      
    }
  }

It was working but after a while, I realized that I need more and more custom logic and additional resolvers in my mesh project and it started to make my mesh project more complicated and hard to maintain as I have more than 20 custom resolvers now. On the other hand, I already have a Nest.js project that contains many services and modules and helper functions that I can reuse for writing my additional resolvers but I realized that I have to somehow duplicate many of those codes inside the mesh project and additional resolvers to write my new additional resolvers.

My question is:
Most of my additional resolvers are like the above sample. Mixing some data sources input or outputs and returning some values and I highly use the SDK version of Mesh through the context parameter in the resolvers.
So as a first step of improvement:
I started to move my Mesh project inside my Nest.js project to make it easier to reuse those existing codes in my additional resolvers. I follow up on that in another question here: #5594

But then I realized that it's not an easy task to do as I couldn't find some information or docs about Mesh + Nest.js integration.
Now as a second solution, I'm thinking about moving my custom logic and additional resolvers to a separate Nest.js project as a new subgraph and implementing them with Mesh SDK, and then adding them to the Mesh as a new data source rather than passing them with additionalResolvers config in meshrc file. In this way, I don't need to integrate the Mesh and my Nest.js project directly but I can link them as a new data source. I can keep my Mesh project clean and easier to maintain and I can use the default fast HTTP handler of Mesh as @ardatan described in this question: #5593

But I'm not sure about the pros and cons of this approach. Do you have any idea about this approach?
I don't want to mix my setup and make that complex in the wrong way I just want to reuse existing codes inside my Nest.js project in my additional resolvers in the Mesh. This is what I mean:

[mnlbox]

@ardatan any idea on this?