Github user LantaoJin commented on the issue:
https://github.com/apache/spark/pull/20574
Thank you Sean for the comments. There are three reasons to do that:
1. Help developer to separate their code from Spark code. I see many
tickets and requests which want to add new pages and finally could not be
accepted, and if we open this extra tab interface, they can just leave their
pages code in a third-part library with a configuration to enable it. No need
merge or hack the Spark code.
2. Another common way is starting a jetty in extra listener and display
pages in this web server. This can work but need driver to launch another
jetty. Also the URL is different with self UI.
3. Use a standalone web server to show spark information. In eBay we did
this things, but we found that almost users still prefer to get information
from origin Spark UI. Although we can hack the code, we still holp Spark can
support this request officially. Thatâs the original purpose of this jira.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
