Detect consumer groups subscribing to several topics using the default “range” strategy, warn the user of possible unbalanced situations
In the producer, it’s now possible to send Avro specific types from a simple JSON representation (bytes as base64, decimal logicalType as number or string, etc.). We’re covering the cases in our documentation.
Improve support of old ksql versions
Support ksqlDB 0.10 and drop previous versions support for queries
Improve how we display the JSON-ish view for logicalTypes not part of unions
Handle older KSQL versions when checking for healthiness
Add new Avro deserializers to work with custom subjects or schemas. It’s also possible to provide a custom schema to deserialize the data and bypass the Schema Registry.
Cleanup and improve Consumer layout to prepare for more options
Fix buggy columns resizing that can happen when refreshing topics view
Improve error message when connection to Kafka fails
Add Smart Groups feature for topics and schema registry. It’s now possible to cleanup long names visually and create dynamic columns based on parts of the names.
It’s now possible to filter Schema Registry Subjects by schema ID. This can return multiple results, when the same schema is used in multiple subjects.
ConsumeFrom configurations are now saved per topic. This will prevent the users to always select the same thing over and over for the same topic. (as we do with the serde and jq pattern)
Shows the partitions progression when consuming a topic.
The max offsets are regularly fetched in the background, for topics ingesting new data in real-time.
The layout of the consume popup buttons has been slightly changed to adapt to the new partitions feature.
In the Consumer Group view, we’ve added a “(must be inactive)” for buttons that are only active when the consumer group is inactive (Reset Offsets or Delete), for the user to know.
We built a new model around the Schema Registry data (json is automatically migrated) to structure the properties (previously in additionalProperties) and display auth methods properly.
Connectors: Always take into account the fields which have changed even when we fallback on the default value. This is to prevent some bug in connectors which are not taking their default value into account. (we don’t send default value, we are relying on what the connector is telling us)