Cool. Exposing metrics through REST api and/or JMX sounds interesting! And thanks for sharing slide and how the event was as well as feedbacks.
moon On Fri, Jul 22, 2016 at 1:10 AM DuyHai Doan <doanduy...@gmail.com> wrote: > Here are the slides for the talk last night at Spark London meetup : > http://www.slideshare.net/doanduyhai/spark-zeppelincassandra-at-synchrotron-64253321 > > There were > 80 people last night and a lot of question about Zeppelin. > > One recurrent remark about Zeppelin is that when running the server on a > long run (several weeks) users are facing instability e.g. sometimes the > server just hangs so they need to restart it. > > Possible improvements could be: > > 1) Expose REST API to return metrics about the state of ZeppelinServer and > of each Interpreter > 2) Expose JMX metrics about the state of ZeppelinServer and of each > Interpreter. Although I prefer REST API, JMX is the de-facto standard for > monitoring and there are a lot of existing monitoring tool that are > designed for JMX > > When I have some bandwidth, I can push some PRs for those monitoring stuff > > > > > > On Wed, Jul 6, 2016 at 4:49 PM, moon soo Lee <m...@apache.org> wrote: > >> Thanks for sharing great event. >> >> >> On Wed, Jul 6, 2016 at 7:20 AM DuyHai Doan <doanduy...@gmail.com> wrote: >> >>> Hello Zeppelin fans >>> >>> If you're living in the London area, I'll be giving a talk about using >>> Spark/Cassandra and Zeppelin to store, aggregate and visualize particle >>> accelerator metrics: >>> >>> http://www.meetup.com/fr-FR/Spark-London/events/232355493/ >>> >>> Also if you know any event related to Zeppelin, please ping me so I can >>> add it to the community page here: >>> https://cwiki.apache.org/confluence/display/ZEPPELIN/Community >>> >>> Regards >>> >>> Duy Hai DOAN >>> >> >
