Hey is the

dbt_models

table meant to continuously add and remove tables? Depending on when we query it during the day sometimes our models are there and sometimes they are completely gone

@Ella Katz This seems like the same behaviour that we're experiencing.

I think I've figured out what the problem is. We have a 3 dbt project setup: •

bronze project

•

silver project

•

gold project

the

silver project

imports the

bronze project

and the

gold project

imports the

silver project

(and in turn also

bronze project

) but it's a one-direction import so the

bronze project

has no idea of the

silver project

and

gold project

and the

silver project

has no idea of the

gold project

we run elementary on-run-end hooks for all three projects towards the same dataset. So my best guess so far is that: • when we run models in the

bronze project

the on-run-end hooks will overwrite the

dbt_models

table with only models from the

bronze project

• when we run models in the

silver project

the on-run-end hooks will overwrite the

dbt_models

table with only models from the

bronze project

and

silver project

• when we run models in the

gold project

the on-run-end hooks will overwrite the

dbt_models

table with only models from the

bronze project

and

silver project

and

gold project

so depending on which project was run last we wither get all models or only a subset. It would be nice if there's a solution to this, it's really helpful for us to have everything in the same project (also for our edr reports and monitoring) but we don't want to run everything from the

gold project

Hi @Kasper Ramström, Yeah that's exactly what would happen if you use multiple envs writing to the same elementary schema. We have an open Github issue about it and we are looking into this. @Sindre Grindheim do you think this is caused by the same issue in your case?

@Ella Katz No, we write to separate environment specific elementary schemas, so that seems like a different scenario.

OK continuing on your thread then 🙏🏼