lxhjava的博客

1.梳理druid预聚合逻辑 摄取预聚合主逻辑： KafkaRecordSupplier.poll() --> IncrementalPublishingKafkaIndexTaskRunner.getRecords() --> SeekableStreamIndexTaskRunner.getRecords() --> StreamAppenderatorDriver.add(record) --> BaseAppenderatorDriver.append() --> ..