[jira] [Updated] (KAFKA-5158) Options for handling exceptions during processing

["Eno Thereska (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/KAFKA-5158?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Eno Thereska updated KAFKA-5158:
--------------------------------
    Issue Type: New Feature  (was: Sub-task)
        Parent:     (was: KAFKA-5156)

> Options for handling exceptions during processing
> -------------------------------------------------
>
>                 Key: KAFKA-5158
>                 URL: https://issues.apache.org/jira/browse/KAFKA-5158
>             Project: Kafka
>          Issue Type: New Feature
>          Components: streams
>            Reporter: Eno Thereska
>             Fix For: 0.11.1.0
>
>
> Imagine the app-level processing of a (non-corrupted) record fails (e.g. the user attempted to do a RPC to an external system, and this call failed). How can you process such failed records in a scalable way? For example, imagine you need to implement a retry policy such as "retry with exponential backoff". Here, you have the problem that 1. you can't really pause processing a single record because this will pause the processing of the full stream (bottleneck!) and 2. there is no straight-forward way to "sort" failed records based on their "next retry time".



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)