Skip to main content
Version: 2.13

Ververica Platform 2.13.2

Release Date: 2024-07-25

Changelog​

Ververica Platform 2.13.2 supports the following versions:

  • Apache Flink® 1.19
  • Apache Flink® 1.18
  • Apache Flink® 1.17

Ververica Platform 2.13.2 supports Apache Flink® 1.19, Apache Flink® 1.18, and Apache Flink® 1.17 under SLA.

For Stream Edition the following Apache Flink® Docker images are available. Please check Ververica Platform Docker Images for all available Apache Flink® images and additional tags.

  • 1.17.2-stream3-scala_2.12-java8
  • 1.17.2-stream3-scala_2.12-java11
  • 1.18.1-stream3-scala_2.12-java8
  • 1.18.1-stream3-scala_2.12-java11
  • 1.18.1-stream3-scala_2.12-java17
  • 1.19.0-stream1-scala_2.12-java8
  • 1.19.0-stream1-scala_2.12-java11
  • 1.19.0-stream1-scala_2.12-java17

For Spring Edition the following archives are available:

Additionally, the Apache Flink 1.16.3 has been patched to fix the FLINK-35565 bug:

Improvements​

Improved Performance for Deployments Page​

We have optimized the deployments page, reducing load times and the overall responsiveness of the platform. Users will now experience faster access to their deployment list, whereas previously, especially in environments with many deployments, page load could take up to one minute.

Improved Performance for Ververica Platform UI Catalog Tree​

Enhanced the catalog tree load times to improve overall responsiveness. Performance improvements were made to ensure that even with a large number of tables and scripts, the catalog tree loads efficiently. This fix ensures a smoother experience when using the SQL Editor and browsing the catalog.

Improved Logging for Kubernetes Operator​

We have refined the Kubernetes Operator logging to provide clearer error messages for CRD deployments. Previously, the operator's logs did not provide sufficients detail for CRD deployment failures, especially for API error cases. This update improves the JSON error output with more detailed error messages written directly to the operator logs, improving troubleshooting and addressing deployment issues.

Improved User Guidance on SQL Editor Preview functionality​

Previously, an unnecessary pop-up info message disrupted the user flow. Now, a warning alert appears below the SSL switch stating, "SQL Editor preview Session Cluster currently does not support SSL communication," and a tooltip icon (ⓘ) near the SQL Editor Previews label provides the same information on hover. This change reduces disruption and enhances clarity for users configuring their sessions.

Enhanced Security for Secret Values in REST API Responses​

Improved the handling of secret values in our REST API. Secret values are now omitted from API responses, matching the Ververica Platform UI behavior where only the secret name is shown.

Improved Cleanup of Dangling Job Artifacts in Blob storage​

Enhanced the Job Service to ensure internal job files are properly removed from the artifacts folder in blob storage. Previously, internal job files were not always removed from blob storage after being deleted from the Ververica Platform database, leading to unnecessary storage usage. This update improves the cleanup service by ensuring all job files under the directories deployments/${name}/jobs/${id}/connectors and deployments/${name}/jobs/${id}/jobgraph are deleted when jobs are marked for deletion.

Upgraded all following Flink connectors to ensure compatibility with Flink 1.19 as they were provided by the community:

  • Flink JDBC connector to 3.2.0-1.19
  • Flink Kafka connector to 3.2.0-1.19
  • Flink Kafka SQL connector to 3.2.0-1.19

Bug fixes​

Session Cluster Deployments stuck in TRANSITIONING State​

Fixed a bug where jobs targeting a session cluster in Ververica Platform would get stuck in the TRANSITIONING state instead of failing when encountering an issue. The fix now treats an HTTP 400 Bad Request response as unrecoverable, allowing the job in session mode to fail early, preventing it to get stuck in the TRANSITIONING state.

Resolved an issue where Flink KafkaSource batch jobs would enter an infinite loop after resetting the offset to an empty Kafka topic. A Flink batch job configured to consume events from a Kafka topic using a specific starting offset would get stuck in an infinite loop if the starting offset was less than the current offset and the topic was empty. The job continually attempted to fetch events without detecting that there were none.

This update ensures the Flink application properly identifies when there are no events to process after an offset reset and gracefully finishes the job, preventing the infinite loop. We reproduced the issue on Flink 1.16.3 with the respective Kafka connector. We have tested and applied a fix, not yet published in the open source connector, that ensures the job is completed as expected without an infinite loop.

  • Upgraded to pyarrow 17 to fix CVE-2023-47248
  • Upgraded to protobuf-java-util 3.23.2 and grpc-protobuf to to fix CVE-2022-3171, CVE-2021-22569
  • Upgraded to flink-connector-kafka 3.2.0-1.19, which fixes CVE-2023-44981
  • Upgraded to flink-connector-jdbc 3.2.0-1.19, which fixes CVE-2022-45868, CVE-2024-1597
  • Dropped bcprov-jdk15on which resolves CVE-2024-30172, CVE-2024-30171, CVE-2024-29857, CVE-2023-33202, CVE-2023-33201

Upgrade​

As always, we recommend upgrading via Helm using the following commands:

$ helm repo add ververica https://charts.ververica.com
$ helm repo update
$ helm upgrade [RELEASE] ververica/ververica-platform --version 5.9.2 --values custom-values.yaml