
I like the way HT does error handling. When creating a custom endpoint for a sync, like a cloud function/lambda/etc, I can define how HT will deal with exceptions. If I surface them to HT one way, HT will automatically retry that sync (if I configure the sync to do that). If surfaced a different way, there will just be a message associated with that record, and the sync will not be retried. This has been helpful in a recent project where many syncs need to be constructed and tested without spamming me with nuisance sync errors. Review collected by and hosted on G2.com.
My typical development pattern is to create a model in HT using straight SQL, build a sync, test, and later create a dbt model. It would be nice if I could easily transition the existing sync to the new model. But since a sync is a child of a model, I have to create a new sync. The "Edit as JSON" button is useful for this, but it would be ideal to just swap the model without rebuilding the sync. Review collected by and hosted on G2.com.