Annie Liang

Currently, Kafka source connector is using changeFeedProcessor to pull changes. Will raise few challenges: 1. Offset tracking 2. load balancing ChangeFeed pull model will be a better suite.

Add benchmark tests for kafka

Sub-partition support in sink connector

Add support so customer can choose upsert vs insert vs etc

Currently, if we enable bulk mode for sink connector, then it will use mode bulk to ingest data to all container, investigate whether we should enable bulk support per container

fixStackOverflowErrorWhenSerializingCosmosDiagnostics

3

**Issue:** Within java SDK, when SDK sending the request to server, it will also spin up a separate task which will cancel the request if SDK did not get the...

fixShowTablesOrShowDatabasesOnlyReturnInCorrectResult

2

**Fixed issues:** - When there are multiple databases under the same account, SHOW DATABASES IN cosmosCatalog only return one database - When there are multiple containers under the same database,...