• Added support for using UNLOAD when bulk syncing from Redshift.

• Redshift Serverless connections could report IAM token expiration during sync exeuction.

• Enforce strict compatibility for integer types in bulk syncs.

Features

• Model syncs with join push-down enabled did not correcly fall back when an error occurred in push-down.
• Switching destinations for model syncs could leave the UI in a bad state.
• Adding columns to Azure/MS SQL Server destinations resulted in an error.

• Fallback to information schema for Redshift tables if SVV_ALL_TABLES is not available.

• Fallback to information schema for Redshift fields if SVV_ALL_COLUMNS is not available.

• Redshift bulk syncs paused reading while data segments were staged.

• NUMERIC/DECIMAL fields with a scale of 0 were incorrectly classified as integers.
• Bulk sync errors which occurred while loading data into Redshift were not surfaced under some circumstances.
• Clarified errors that occur loading destination tables for model syncs.

Fixes