Introduction

Sundeck provides data engineers an entirely new way to work with Snowflake in combination with their favorite existing tools. Data engineers, analysts, and DBAs are now able to capture queries before they are executed, then: reject, reroute and modify them; run pre- and post-queries before or after the original query; and write queries more easily with enhanced SQL syntax.

Several simple data engineering ideas have historically been surprisingly difficult to implement. Want to stop users from accidentally running billion-record queries? Want to introduce custom syntax or hints to make analysts more productive? Want to automatically route queries amongst different warehouses based on the nature of the query or the current load on the system? Want to have per-user consumption quotas? Prior to Sundeck, these things were difficult to do. With Sundeck, these tasks are trivial.

dbt made data transformation open and extensible. Building for this set of principles sparked a creative renaissance in data engineering. Sundeck is applying those same principles to every tool that sends queries to Snowflake. The relationship between tool and data cloud is now open and accessible for data engineers and analysts to adapt as they see fit.

To keep things simple, your Sundeck account includes a powerful default flow which blocks common anti-patterns, provides helpful Snowflake syntax enhancements, enables granular cost control, and more … with minimal or no configuration.

For users who want to go deeper into hands-on query engineering, Sundeck provides a collection of query hooks which can be used to extend the out-of-the-box flow, or can be used to compose entirely new flows. Hooks can intercept queries and take actions on the queries before (or after) they execute in Snowflake. The decisions of when to intercede and which hooks to execute are entirely configurable by data engineers. Hooks fire based on conditions including the nature of the query, the tables involved, and the user submitting the query. Specific hooks include operations like expanding or rewriting the query text, executing a secondary Snowflake query to evaluate advanced quotas, rejecting the query due to policy and routing a query to an alternative warehouse.

Sundeck is built to work just like Snowflake. That means all administration can be done with SQL and any tool that can talk to Snowflake also works with Sundeck. Additionally, Sundeck’s unique private results path technology enables data engineers to block or enhance queries while ensuring results stay within an organization’s boundaries. When combined with Sundeck’s SOC 2 certification, organizations can be assured that Sundeck is safe.