+1 Vinoth Chandar <vin...@apache.org> 于2020年4月22日周三 下午2:24写道:
> +1 from me as well > > On Mon, Apr 20, 2020 at 9:37 PM vino yang <vinoy...@apache.org> wrote: > > > Hi Raymond, > > > > Thanks for opening this discussion. > > > > IMHO, as Hudi's user base grows, we need to enhance our metrics reporter. > > From an ecological point of view, this is also very important. > > > > So, +1 from my side. > > > > Best, > > Vino > > > > Shiyan Xu <xu.shiyan.raym...@gmail.com> 于2020年4月21日周二 上午10:59写道: > > > > > Hi all, > > > > > > I'd like raise the topic of supporting multiple metrics reporters. > > > > > > Currently hudi supports graphite and JMX. And there are 2 proposed > > reporter > > > types: CSV and Prometheus > > > https://jira.apache.org/jira/browse/HUDI-210 > > > https://jira.apache.org/jira/browse/HUDI-361 > > > > > > I think supporting multiple metrics backends gives Hudi competitive > > > advantage on user expansion. It reduces the friction for different > > > organizations to adopt Hudi. And we only need to support a few popular > > ones > > > to achieve that. > > > > > > In terms of determining the list, as mentioned by @vinoyang, flink has > a > > > nice list of supported ones: > > > > > > > > > https://ci.apache.org/projects/flink/flink-docs-release-1.10/monitoring/metrics.html#reporter > > > which can be used as a reference. > > > > > > From that list, I'd like to propose supporting Datadog as well, due to > > its > > > popularity. May I get +1 on this? > > > > > > Thank you. > > > > > > Regards, > > > Raymond > > > > > >