[email protected] <[email protected]>于2020年6月15日 周一17:58写道：
> How about spilt them? > For event trace data (about job execute timestamp and result), persist into > MySQL or other databases, for metrics, persist into Prometheus. I think separating them will make you can't use Grafana. I think alarm, monitor and observability should be an important and separated topic. Sheng > > ------------------ > > Liang Zhang (John) > Apache ShardingSphere & Dubbo > > > Sheng Wu <[email protected]> 于2020年6月15日周一 下午2:02写道： > > > What event trace data? Prometheus is focusing on metrics only. Only > > statistic data is suitable. > > > > Sheng Wu 吴晟 > > Twitter, wusheng1108 > > > > > > [email protected] <[email protected]> 于2020年6月15日周一 下午1:34写道： > > > > > Hi Sheng, > > > > > > The mean of `Support prometheus storage` is putting the event trace > data > > > info prometheus ecosystem, we can use push gateway to collect the data > > and > > > use grafana to display the result. > > > > > > ------------------ > > > > > > Liang Zhang (John) > > > Apache ShardingSphere & Dubbo > > > > > > > > > Sheng Wu <[email protected]> 于2020年6月15日周一 上午9:26写道： > > > > > > > Hi Liang > > > > > > > > About SkyWalking, that is another topic. It is fine to do that later. > > > > My question is, what do you mean `Support prometheus storage`. > > Prometheus > > > > is a monitoring system, basically. What does it storage you mean? > > > > > > > > Sheng Wu 吴晟 > > > > Twitter, wusheng1108 > > > > > > > > > > > > [email protected] <[email protected]> 于2020年6月13日周六 > 上午1:28写道： > > > > > > > > > For persist job event trace. We can consider about use SkyWalking > to > > > > trace > > > > > the job event, is it possible? > > > > > > > > > > ------------------ > > > > > > > > > > Liang Zhang (John) > > > > > Apache ShardingSphere & Dubbo > > > > > > > > > > > > > > > Sheng Wu <[email protected]> 于2020年6月12日周五 下午4:54写道： > > > > > > > > > > > [email protected] <[email protected]> 于2020年6月12日周五 > > > 下午4:24写道： > > > > > > > > > > > > > Hi all, > > > > > > > > > > > > > > We are beginning to develop ElasticJob, and want to add some > new > > > > > features > > > > > > > in version 3.0. > > > > > > > > > > > > > > I want discuss the feature of it. > > > > > > > > > > > > > > Some ideas for new features: > > > > > > > > > > > > > > - Support one time trigger for ElasticJob lite > > > > > > > - Support job dependency based one sharding item > > > > > > > - Use native ZooKeeper API instead of curator > > > > > > > > > > > > > > Some ideas for new architecture: > > > > > > > > > > > > > > - Refactor Job API, just keep SimpleJob, make other job types > > > > introduce > > > > > > via > > > > > > > SPI > > > > > > > - Redesign Job domain, make job as top level interface, which > > > > composite > > > > > > by > > > > > > > task (sharding item) > > > > > > > - Refactor event trace module > > > > > > > - Split event trace and elasticjob-core module > > > > > > > - Open SPI for event trace data storage > > > > > > > - Support prometheus storage > > > > > > > > > > > > > > > > > > > What is Prometheus storage? > > > > > > > > > > > > > > > > > > Sheng Wu 吴晟 > > > > > > Twitter, wusheng1108 > > > > > > > > > > > > > > > > > > > > > > > > > > ------------------ > > > > > > > > > > > > > > Liang Zhang (John) > > > > > > > Apache ShardingSphere & Dubbo > > > > > > > > > > > > > > > > > > > > > > > > > > > > -- Sheng Wu 吴晟 Apache SkyWalking Apache Incubator Apache ShardingSphere, ECharts, DolphinScheduler podlings Zipkin Twitter, wusheng1108
