What are your top tips on testing cost efficiently in production on big query?


We’ve got loads of tests on all our models, but running all the tests every time we do a production run has started creating costs. We’re in BigQuery. We were considering changing some of our staging models from tables to views, but this would incur even more cost if we’re testing several columns in the view.

Perhaps we have too many tests, but they are very useful when developing to check we’ve not messed something up. (We dev on limited datasets for costs).

Are there nice solutions for setting tests to only run in certain environments? (I know we could tag all the tests so that they are excluded in production - is that the best way?)

Any other tips on how others have covered this?