Are MySQL, Postgres, and SQLAlchemy the only options for Dagster's backend? My company uses SQL Server for production and analytics data, and I was hoping to use it for Dagster metadata as well.

hi @Caelan Schneider! you're correct that those are the only storage backends which are supported out of the box. It is possible to subclass the existing SqlEventLogStorage class to implement a custom backend, although I'll admit that would take a bit of effort.

Are there any plans to officially support SQL Server as a backend for Dagster at any point, perhaps in the near future? Otherwise, would it be possible to simply write in the information for SQL Server in the yaml using the syntax for the MySQL backend for example? Perhaps with different connection string? Or is it more involved than that? The last resort I can imagine, if on really wanted to get the dagster metadata into SQL Server, is to use the SQLite backend and then create a job to ingest that data. How crazy does that sound to you?

Hi @Laurids Mikkelsen! Unfortunately, no plans in the near future to offer official support here. I think the only realistic option here would be to subclass the SqlEventLogStorage class. As mentioned, it would be somewhat involved, but following the basic template of the MySQLEventLogStorage class would likely get you most of the way there

Hi @owen. I have same requirement and willing to contribute. Would it be possible to merge PR even if SQL server support is not on the roadmap?

hi @Nikolaj Galak! yep we'd definitely be happy to accept a contribution of that sort.

@Nikolaj Galak @owen any updates on this?

@Laurids Mikkelsen I don't have clear roadmap for it at the moment. Data stack in the company where I work is based on Microsoft stack which is main driver for enabling MS SQL. This is low priority task as we are happy with MySQL backend but it adds one more application to already complex landscape. MS SQL is widely used in other applications we have so it is natural to implement respective Dagster backend. The plan is to make a solution for MS SQL at some point in September and open souce it to the Dagster community.