Skip to main content

Community

Destination Improvement: add session variables for Snowflake destination

Please sign in to leave a comment.

Comments

3 comments

  • Official comment

    Hello,

    Thank you both for your replies on our feature request portal! We have been evaluating different layers/levels of hierarchy when working with objects like Destinations within Fivetran.

    The prime use case would be if there are multiple teams within the same organization, and different teams might have different warehouses. I can imagine extrapolating this to a point where within a team, there might be different connector groups that use different warehouses.

    Could you help clarify how you would like to use query tags? At Fivetran we have a principle to keep connector configuration to a minimum, so would love to understand your use case and desired behavior with query tags.

    Hello - Upvoting this, not only for better cost analysis, but also optimizing compute power and managing queue times. It would be even better if we could customize Snowflake warehouses at the table level (some connectors have 1 or 2 large tables and the rest are small). 

    Hi Kevin, 

    The main reason I'd like to be able to set up 1 connector and split it to use different warehouses is to maximize cost efficiency on Snowflake and reduce queue times so that the connector can sync faster.

    For example, one of our connectors has a page_events table that is massive (think hundreds of thousands of records added per hour or day and that can grow depending on usage of our product), but almost all the other tables we're pulling in are comparatively small (like a table of customer level information). It would be nice to have the option to point large tables or tables that take a long to copy to larger warehouses and leave small tables on smaller warehouses + have more insight into the efficiency of resource usage by table (I believe that's actually available but a little buried / simple in the UI at the moment).

    I'm not sure if this would be possible for all connectors because some of them don't have the option to select individual tables, but at least for some of the more popular ones with more advanced APIs, like Salesforce, it would be fantastic if we could customize the loading process more when connected to Snowflake. 

Didn’t find what you need?

Contact support