Skip to main content
Kubit connects directly to Snowflake and queries your data in place. You create a dedicated role, user, database and warehouse, all named KUBIT, then grant the role read-only access to the schemas Kubit should see. Nothing is copied out of your account. The setup takes about 20 minutes.

Before you start

  • A Snowflake role that can create roles, users, databases and warehouses (ACCOUNTADMIN, or SECURITYADMIN plus SYSADMIN)
  • openssl on your machine, to generate the key pair
  • The schemas that hold the tables or views Kubit should query
  • If your Snowflake account has a network policy, Kubit’s IP ranges. See Allow Kubit’s IP ranges.

Steps

1

Create a role

Create a role named KUBIT.
2

Generate a key pair

Kubit authenticates to Snowflake with key pair auth. Generate an encrypted private key and its public key. openssl asks for a passphrase. Keep it, since Kubit needs it with the private key.
Snowflake’s key pair authentication guide covers the options.
3

Create a user

Create a user named KUBIT with the public key from the previous step. Paste the contents of rsa_key.pub without the -----BEGIN PUBLIC KEY----- and -----END PUBLIC KEY----- lines.
4

Create a database

Kubit keeps a small auxiliary table in its own database to support some features. Create a database named KUBIT and make the KUBIT role its owner, so Kubit can create and run the tasks that need an owner role.
5

Create a warehouse

Create a warehouse named KUBIT and grant the KUBIT role MONITOR, USAGE and OPERATE on it. Start small. Warehouse configuration explains when to size up.
6

Grant read access

Give the KUBIT role read-only access to each schema that holds the views Kubit queries. The FUTURE grant covers views you add later. Replace <database> and <schema> with your own names.
7

Add the connection in Kubit

Kubit needs four values: your Snowflake account locator, the private key (rsa_key.p8), the key passphrase and the warehouse name (KUBIT).
Snowflake Connection setup form
For help with the setup, share the same four values with the Kubit team through a secure channel such as a password manager.
8

Split test and production (optional)

Create a second role, user and warehouse for development. Keeping development queries off the production warehouse avoids any impact on live workloads.

Verify

Check the grants as an admin. The output lists USAGE and SELECT on your schemas, and MONITOR, USAGE and OPERATE on warehouse KUBIT.

Best practices

Clustering keys

Kubit queries time series data, so cluster fact tables larger than 1 TB by date. Most Kubit reports also filter by event name, so add event name as the second clustering column. Snowflake’s clustering keys reference covers the syntax.

Warehouse configuration

Four warehouse settings affect Kubit’s cost and speed.

Next steps