[jira] [Commented] (CAMEL-3631) Graceful shutdown - Add logic to detect slow/hung in-flight exchanges

["Ashwin Karpe (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CAMEL-3631?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13023418#comment-13023418 ] 

Ashwin Karpe commented on CAMEL-3631:
-------------------------------------

Hi Claus,

No problem. I can store the exchange in the ConcurrentMap. I was hesitating to store the exchange itself due to the runtime implications of doing this.

At runtime,this could require a copy of the exchange to be saved in the Map until the exchange was delivered and could lead to significant memory growth if the exchange payloads were quite large. Can you please provide your thoughts on this.

I could definitely redesign the code to not add more interfaces to the SPI.

Cheers,

Ashwin... 
  

> Graceful shutdown - Add logic to detect slow/hung in-flight exchanges
> ---------------------------------------------------------------------
>
>                 Key: CAMEL-3631
>                 URL: https://issues.apache.org/jira/browse/CAMEL-3631
>             Project: Camel
>          Issue Type: Improvement
>    Affects Versions: 2.6.0
>            Reporter: Claus Ibsen
>            Assignee: Ashwin Karpe
>             Fix For: 2.8.0
>
>         Attachments: Camel-3631.patch
>
>
> We should improved the graceful shutdown to be more verbose and provide more detailed information when it appears in-flight exchange(s) appear to be hung.
> We have the created timestamp on the exchange, and if we let in flight repo, track each exchange, then we can grab the exchange and thus know how long time it currently has been in-flight. And then if the time is > threshold, we can output that in the logs.
> It requires the in flight registry now also keep track of each exchange. We may consider this as an optional feature you should enable, as it requires us to put/remove Exchange into a ConcurrentMap to keep track of each exchange.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira