[ 
https://issues.apache.org/jira/browse/SPARK-23235?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16348758#comment-16348758
 ] 

Imran Rashid commented on SPARK-23235:
--------------------------------------

Just brainstorming I'd like to get some opinions on:

I started to wonder whether this endpoint should be *just* the threaddump, or 
if we should make it richer for collecting even more metrics, other stuff which 
is too expensive / verbose to put in the all executors endpoint.  For example, 
the ids of running tasks and stages, extended memory metrics, or even 
everything available in the executor metric system.  And taking that idea one 
step further -- rather than having this info go via the driver, we could expose 
an endpoint in the executors themselves.  It still wouldn't be super efficient 
to call that, but better than routing this info via the driver.

anyway, my current feeling is that those might be interesting ideas, but in any 
case, we should just add the original proposal for now.  Its simple to do, 
gives parity with the UI, and I don't think it would conflict if we decided to 
do those other things sometime in the future.  But I wanted to see if others 
have any opinions, and I'll mull it over a bit too.

[~jerryshao] you might have some thoughts on this as well.

> Add executor Threaddump to api
> ------------------------------
>
>                 Key: SPARK-23235
>                 URL: https://issues.apache.org/jira/browse/SPARK-23235
>             Project: Spark
>          Issue Type: New Feature
>          Components: Web UI
>    Affects Versions: 2.3.0
>            Reporter: Imran Rashid
>            Priority: Minor
>              Labels: newbie
>
> It looks like the the thread dump {{/executors/threadDump/?executorId=[id]}} 
> is only available in the UI, not in the rest api at all.  This is especially 
> a pain because that page in the UI has extra formatting which makes it a pain 
> to send the output to somebody else (most likely you click "expand all" and 
> then copy paste that, which is OK, but is formatted weirdly).  We might also 
> just want a "format=raw" option even on the UI.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org

Reply via email to