Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 57.69%. Comparing base (f50af60) to head (d135f04).

@@           Coverage Diff           @@
##             main     #675   +/-   ##
=======================================
  Coverage   57.69%   57.69%           
=======================================
  Files          48       48           
  Lines       33079    33080    +1     
=======================================
+ Hits        19084    19085    +1     
  Misses      13995    13995           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

LGTM. Docs?

I couldn't find any docs for the other related fields, e.g. cluster_id, so i think we are not documenting it (at least yet). @mgyucht please lmk if documentation is needed

Config doesn't have to validate the value in serverless_gc_id, this will be done in the downstream applications.

Isn't the only allowed value at the moment "auto" then? If this is the case, shouldn't it have validation?

Config doesn't have to validate the value in serverless_gc_id, this will be done in the downstream applications.

Isn't the only allowed value at the moment "auto" then? If this is the case, shouldn't it have validation?

i'm open to both options, the pros for not validating in config IMO are:

1. config is doing little validation for other fields and mostly leaves validation to the downstream applications. not sure if we should single out serverless_gc_id field
2. if dbconnect or any other downstream applications will need to add another value except auto, databricks-sdk code doesn't need to change and the user has one less package to upgrade

Sorry for being horrible annoying. According to https://docs.databricks.com/en/compute/serverless.html I would just name the field serverless_compute_id.