The process for configuring a Snowflake connection in Kraken is identical, regardless of the cloud platform on which Snowflake is running.
- Google Cloud
Setting up Kraken to connect to your Snowflake environment requires a user that has read privileges for using training data and apply data in Kraken. If you want Kraken to write predictions to Snowflake, the user needs drop, create, read and write access to the Snowflake table(s) where Kraken predictions will be written.
If your Snowflake database is not accessible outside of your network, you may need to contact your IT department to whitelist Kraken's IP address. Please contact your friendly neighborhood Big Squid Customer Success Manager for more information.
Adding a New Snowflake Connection in Kraken
- Log into Kraken
- Click the "hamburger menu" in the upper left
- Click on "Manage Data"
- Click on the Snowflake data provider card and enter your credentials, then click "CONNECT"
Account: Specifies the portion of the account specific to your Snowflake environment. Do not include "https://" prefix or ".snowflakecomputing.com" suffix since Kraken automatically adds these when establishing the connection.
AWS example: abcdefg-mycompany_myregion
If your Snowflake environment is in the AWS US West region, you do need to specify the region (e.g. abcdefg-mycompany)
Username: Specifies the login name of the Snowflake user to authenticate
Password: Specifies the password of the user that is connecting to Snowflake
Database: Specifies the default database for Kraken to use.
Schema: Specifies the default schema for Kraken to use.
Warehouse: Specifies the default warehouse Kraken should use to run queries
Role: (OPTIONAL) Specifies the role for Kraken to use with the associated username.
Video: Kraken + Snowflake Integration Overview
- Instructional overview of connecting Kraken with Snowflake
- Information on pulling training and apply data from Snowflake into Kraken
- Explanation of how to synchronize predictions from Kraken to Snowflake
- A brief overview of Automated Prediction Download functionality