Hello! I want to track my dbt models execution time, to find out if a model is suddenly taking longer time and so costs more. it doesn’t have to be an increase in volume, but even changes to the model’s sql that can cause that. I was thinking of using an anomaly detection test on the

elementary.dbt_run_results

table, but I need a test that will check the average execution time for every model, which I don’t think is available. any chance you have a solution or a recommendation how to approach this?

Hey @Anat Ezra, have you tried to use

where_expression

with this test ?

it’s more of a

column_anomalies

test of type average, combined with the

dimension_anomalies

tests, in the sense of a dimension split. I thought about using

columns_anomalies

with

where_expression

, which will require a test for every model value, which is not ideal

oh ya, columns_anomalies will work 🙂

do you happen to know if I can use a for loop to overcome this? I know jinja is limited in yaml files in dbt

you want to loop over all available models ?

theoretically (I will probably want to limit this to a reasonable list of models)