When AppSync subscription payloads are defined by the initiating mutation instead of what I ASKED for, and filtering only works when the initiating mutation specifically queries for that field in the respond payload that it doesn't even need.

The only reason to migrate to AppSync is the fact that a managed Apollo Server is nowhere to be found.

You heard me, Apollo? Come on, even MongoDB got Atlas!

Every once a while, I secretly wished AppSync will be equipped with persisted queries and batch queries.

Was I not praying hard enough?

Someone in GitHub made a good pose.

A paid solution like AppSync should be complete and does not require hacking, using the AppSync JavaScript client is just a pain in the ass.

The entire AppSync/Amplify SDK for android is a shitfest. The support engineers don't know what the devs are doing, and the devs don't give 2 shits. It shouldn't take 10 fucking hours to configure an existing API and Auth to work with an app and then run into issues with the code that the fucking SDK generates. Fucking buffoons

I'm looking for advice...

Has anyone experience with the AWS cloud?
I'm arguing with my future company partner about it. He's totally old school but is responsible for the server stuff... He does the backend for a urgently needed webapp and it takes so long (he still works in his old company the next months).

The frontend (my part) is nearly ready. I could work on the backend fulltime, but I would choose AWS Appsync with offline sync etc. First it would be quick and dirty, because it's really urgent.. he wants to do all super perfect...

How can I handle that? I talked to him many times about that, but he always says it should be done right and takes time. but for me, it's to much time. The webapp is relatively small and the work now already takes about 2 or 3 months..