06.15.2023 Release Notes

These are the release notes for June 15, 2023

:sparkles: ENHANCEMENTS :sparkles:

  • All environments (Gen1/Gen2)
    • Several enhancements are now available for Read Connectors for cloud blob storage services (Amazon S3, Azure Data Lake Storage, Google Cloud Storage).
      • ZIP files are now supported! You can also use regular expressions to filter by file name within ZIP files, when there is more than one file within a ZIP file.
      • For Custom Python Parsers, you can add an optional infer_schema() method, which gives you control over the types returned by the parser during schema generation.
    • Performance of schema generation in the CData Read Connector has been optimized.
    • The Apache Iceberg library has been upgraded to 1.3.0.
  • Gen2 environments
    • Several improvements have been implemented that help address cases where a component is "stuck in ready".
      • Show Component metadata operation state in the User Interface (UI).
        • Ascend was performing backend metadata operations that were not visible to users in the UI and made components look like they were "stuck in ready".
        • This enhancement adds new states to the UI (including "Waiting to Update", "Updating Metadata", "Waiting to Sweep" and "Sweeping"). These are now documented in our component state documentation.
      • Internal platform change to separate processing job tasks from metadata operations.
    • Snowflake Data Plane
      • The Warehouse Override option "Metadata Warehouse" has been renamed to "Interactive Warehouse" in the Snowflake Data Plane configuration.
        • This is to give customers the option to separate some types of query traffic from data pipeline processing query traffic.
      • Query traffic that is routed to the Interactive Warehouse now includes:
        • Queries that enable the records preview tab in components;
        • Data Plane usage metric gathering queries run in the backend;
        • Snowflake schema inference jobs.
      • User queries run through the Ascend Query tab are still sent to the "Default" warehouse.
      • Users that actively use the existing ingest_warehouse and metadata_warehouse settings for Data Share components (via the Ascend Python SDK) should change their code to use the new warehouse_usage_overrides field (see docs).

:wrench: BUGFIXES :wrench:

  • All environments (Gen1/Gen2)
    • Use binary read/write functions to avoid changing newlines on user code during SDK download/apply on Windows.
      • Fix was released in Ascend SDK v0.2.59.
    • Fix a bug where some event fields used in a Webhook Notification template were not being populated.
      • Event fields now being populated include: event.component.name, event.component.type, event.component.uri.
    • A condition that would occasionally result in missing Billing Usage information in the Observability tab has been handled.
  • Gen2 environments
    • Fix a bug where using the Ascend Python SDK to download a Data Share definition resulted in an empty list of Data Services with access to the share.
      • This triggered a bug where the "hidden_from_host_data_service" attribute was being inadvertently set to True. This caused the Data Share to be hidden from the data service it was published in.
      • Fix was released in Ascend SDK v0.2.60.
    • Snowflake Data Plane
      • Some cases of components stuck in Running, Listing and Ready states have been fixed by improving Data Plane Snowflake connection management and more eagerly closing connections.