I've been debating taking a old (2yrs+) project to .NET Core, from .NET Framework 4.5.2.

(Edit: This project has been idle for a few years. It was a startup I worked on for 18mo, but has been just sitting there. I did upgrade the code to 4.5.2 last year to keep it somewhat current. If I stick with .NET Framework, I'll bring it up to 4.7.2 (?))

I'm going to be pivoting the project to be the back-end for a new mobile app (Xamarin).

The current project consists of ASP.NET MVC web services, Service Fabric services, and leverages Azure DocumentDb (now CosmosDb) and Service Bus. Everything was deployed to Azure, obviously.

I've been following .NET Core from the early days, and while I like the potential of non-Windows hosting using Docker containers, I've already seen some limitations in .NET Core-compatible packages, i.e. Service Bus management in .NET Core client (topic subscriptions are missing).

With this new project, I'm planning on taking more advantage of the graph db capabilities in CosmosDb, and creating a GraphQL based API. Also, I'll likely leverage Azure Functions.

As much as I'd like to be on the latest-greatest, I'm interested to get some opinions if making this port is just wasted work for little value, or I'd be setting myself up better for the future?

Much appreciated!