Create static-time-partitioned jobs

[javiermas]

Following the documentation of dagster, imagine you have a job that's partitioned by continent and time. For example, you want to process every continent's data monthly, but separately (each continent has a different configuration).

I believe there is no way to model this explicitly at the moment. Personally, what I've ended up doing is creating a time-partitioned job that reads the config for all continents and then processes their data independently via DynamicOutputs. I don't like this way of modeling the ETL: it forces the chaining of ops and it creates an implementation that is very dagster-specific.

Jenkins had a feature that could serve as inspiration, configuration matrices