You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@zeppelin.apache.org by Spico Florin <sp...@gmail.com> on 2018/05/01 08:04:31 UTC
How to track a zeppelin job for multiple job submission request
Hello!
I have a zeppelin notebook that I would like to be exposed as a REST
service to multiple users.
A user can request multiple times the results from the REST service backed
by zeppelin.
I would like the calls to service to be asynchronous and to use the async
api
https://zeppelin.apache.org/docs/0.7.3/rest-api/rest-notebook.html#run-all-paragraphs
or
https://zeppelin.apache.org/docs/0.7.3/rest-api/rest-notebook.html#run-a-paragraph-asynchronously
The calls are performed from a web browser. The results from the service
are put in a kafka topic and sent back to the client via web socket.
The flow is like this:
User (web browser)->Zeppelin REST job (use Spark interpreter)->Kafka
topic->Socket.io(kafka web socket plugin)->Web browser
Due to the fact that the mentioned REST API doesn't provide the job id:
- how can I successfully return back the results to the client requester?
-how can I implement such a described workflow?
-how can I distinguish between client requests?
I look forward for your answers.
Best regards,
Florin