Known Issues
Before installing or upgrading AtScale, please review this section to be aware of all known issues.
For models with more than 1600 attributes, AtScale fails to answer all PostgresSQL metadata queries from BI tools
When connected to BI tools via the PostgresSQL interface, AtScale fails to answer all metadata queries on models that have more than 1600 attributes. This is due to a limitation of the PostgresSQL database used to answer these queries.
ATSCALE-22911
Aggregates removed on redeploy
In the new deploy mechanism, when the AtScale UI compiles SML into a project.xml
, it is reinitialized, causing aggregates to be invalidated. This results in individual aggregates being rebuilt after each deployment.
ATSCALE-22187
Multiple Databricks connections are unsupported
AtScale 2024.7.0 does not support more than one Databricks connection per data warehouse configuration. Assigning an AtScale engine (large, small, or system) to a secondary connection does not save. This issue will be resolved in a future release.
ATSCALE-22049
Databricks clustered tables are unsupported
Databricks clustered tables contain hidden columns, which cause two issues in AtScale:
- They interrupt Data Preview for the datasets.
- Their presence in the datasets' SML causes model deployment to fail.
As a workaround, you can remove these columns from the datasets' underlying SML. This will enable you to successfully deploy your models.
This issue will be resolved in a future release.
ATSCALE-22036
Multi-Model Catalogs Accessed via the Inbound PostgresSQL Interface not Completely Updated on Deploy
For AtScale installations that use the PostgresSQL interface, deploying catalogs with multiple models can create race conditions that result in only a subset of those models being created in the metadata database.
Due to this issue, customers who make inbound SQL connections using the PostgresSQL interface AND who deploy catalogs with multiple models should not upgrade to this release.
ATSCALE-22003
Cube Overrides Interface unavailable
In installer-based AtScale, you can override engine settings at the data model (cube) level. In containerized AtScale, the ability to set these overrides is currently only available via API. A new mechanism for data model override management will be introduced to containerized AtScale in coming releases.
ATSCALE-17458
Snowflake Key File authentication method unavailable
For Snowflake data warehouse connections, the Key File authentication method is currently unavailable. All other methods are functional. The Key File method will become available in a later release.
ATSCALE-21397
AI-Link
AtScale AI-Link is not currently supported for this release. This will become available in a later release.
Services and Integrations page not available
The Services and Integrations page in Settings is currently unavailable. This will become available in a later release.