hudeqi opened a new pull request #10716: URL: https://github.com/apache/kafka/pull/10716
1. besides `latest` and `earliest`, we also add `nearest`: reset to either latest or earliest depending on the current offset (i.e. this policy won't trigger under the scenario when we see a partition for the first time, without committed offsets; it will only trigger for out-of-range). 2. `latest-on-start`, `earliest-on-start`: reset to either latest or earliest only when we see the partition for the first time without committed offset; when out-of-range default to `none`, i.e. throw exception. 3. an additional `limitTimeStamp` limit used for `latest/earliest/latest-on-start/earliest-on-start`: it means we only reset to latest / earliest if its partition's first record timestamp is smaller / larger than the given `limitTimeStamp` parameter, otherwise, reset to earliest / latest. set the `limitTimeStamp` value to the consumer group started timestamp, when new partitions are added it would reset to `earliest` to avoid losing data. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org