[jira] [Commented] (NIFI-9581) PutElasticsearchRecord should make it easier for users to distinguish between Records that resulted in Elasticsearch errors and those that did not

["Chris Sampson (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/NIFI-9581?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17477420#comment-17477420 ] 

Chris Sampson commented on NIFI-9581:
-------------------------------------

{{PutElasticsearchRecord}} and {{PutElasticsearchJson}} should also emit {{SEND}} Provenance Events to indicate that data is sent to Elasticsearch.

> PutElasticsearchRecord should make it easier for users to distinguish between Records that resulted in Elasticsearch errors and those that did not
> --------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: NIFI-9581
>                 URL: https://issues.apache.org/jira/browse/NIFI-9581
>             Project: Apache NiFi
>          Issue Type: Improvement
>    Affects Versions: 1.14.0
>            Reporter: Chris Sampson
>            Assignee: Chris Sampson
>            Priority: Minor
>
> The PutElasticsearchRecord processor will currently optionally log Records from the input FlowFile that were marked as resulting in an {{error}} within Elasticsearch; these same Records can optionally be directed to the {{errors}} Relationship from the Processor if the user provides an appropriate {{Error Record Writer}}.
> The original {{input}} FlowFile is always sent to the {{success}} Relationship irrespective of whether any/all Records resulted in Elasticsearch {{errors}} (unless the call to Elasticsearch itself fails, in which case the FlowFile is routed to {{failure}} or {{retry}} depending upon the Exception type).
> It would be more useful if the Records that did not result in {{errors}} were also written to a Relationship such that users could process those separate to the errored Records without having to try and find them from the original {{input}}. This would more closely mirror what is already done as part of the {{PutElasticsearchHttpRecord}} processor, which separates Records between the {{success}} and {{failed}} Relationships depending upon the Elasticsearch responses for each Record.



--
This message was sent by Atlassian Jira
(v8.20.1#820001)