Kylin use Kafka consumer API to retrieve the data and save the offset into
metadata when the cube is built successfully. There will no data lost. If
the build failed, you could re-trigger the build again.

The data(streaming cube) available latency quite depends on your cluster
capability. In some existing demos, we found the latency could be 3-5
minutes. If you want process more data, please extend your cluster and
check the result.

在 2017年4月12日 下午4:52,Chen Xinxin 陈昕鑫 <[email protected]>写道:

> 想请问一下,在麒麟2.0的版本里,对从kafka读取实时数据支持的程度,
>
>
>
> 1,  对丢失数据的改进,在一秒里多大数据量的情况下,丢失数据的概率。
>
> 2,  能支持多大时间粒度的实时查询
>
> 3,  在保证麒麟实时计算准确的情况下,最大支持的数据量和qps
>
>
>
> 谢谢能给予及时回复
>
>
>
>
>
>
>
> *Thanks&Best Regards!*
>
> *陈昕鑫** Chen Xinxin*
>
> Mobile: +86 18918133616 <189%201813%203616>
>
> E-mail: [email protected]
>
> Website: www.fosun.com
>
>
>
> 复星集团云济信息科技有限公司 | 中国上海市徐汇区宜山路1289号B幢11楼 邮编:200233
>
>
>

Reply via email to