cube-js/cube

Cubestore - ADD Snowflake Storage Integration to Access Amazon S3

Open

#5,472 opened on 2022年10月17日

GitHub で見る
 (1 comment) (0 reactions) (0 assignees)Rust (1,965 forks)batch import
driver:snowflakehelp wanted

Repository metrics

Stars
 (19,563 stars)
PR merge metrics
 (平均マージ 5d 16h) (30d で 138 merged PRs)

説明

Is your feature request related to a problem? Please describe. This is barely documented at all, but it seems Snowflake driver's S3 Export Bucket functionality currently works by:

  1. Creating temp pre-agg tables in Snowflake (with CREATE commands)
  2. Using a Snowflake COPY command to move this temp data into the S3 bucket you've specified in the CUBEJS_EXPORT_ and CUBESTORE env variables, including the access key/secret.

This is not the ideal way to integrate Snowflake with S3. Snowflake has a concept of an S3 Storage Integration.

Describe the solution you'd like Users should be allowed to set this up and specify their integration, rather than need to pass the access key/secret from Cube to Snowflake. There are many preferred ways to access an S3 bucket from a certain place (for example, EKS where Cube may be running), and the underlying Rust-AWS-auth lib seems to already support them.

Matter of fact, this Snowflake integration seems to have already been added for GCP buckets.

Describe alternatives you've considered No workarounds that are secure.

コントリビューターガイド