[jira] [Created] (ATLAS-4348) Atlas-Kafka Hook : When a producer publishes messages to multiple topics, the latest relationship is marked ACTIVE , rest are marked DELETED

["Radhika Kundam (Jira)" <ji...@apache.org>]

Radhika Kundam created ATLAS-4348:
-------------------------------------

             Summary: Atlas-Kafka Hook : When a producer publishes messages to multiple topics, the latest relationship is marked ACTIVE , rest are marked DELETED
                 Key: ATLAS-4348
                 URL: https://issues.apache.org/jira/browse/ATLAS-4348
             Project: Atlas
          Issue Type: Task
          Components:  atlas-core
    Affects Versions: 2.1.0
            Reporter: Radhika Kundam
            Assignee: Radhika Kundam


*Steps to reproduce the issue*:

Created 2 topics : test1, test2

Using console consumer produced messages to test1 and test2.

Relationship between console-producer-lineage and test1 is ACTIVE now.

Relationship between console-producer-lineage and test1 is set to DELETED when messages are published to test2.


*Root cause of issue:*

When messages published on topic, Atlas hook receives partial_update request with only one topic as output and by default partial update deletes relationships between previous topics by considering the only topic(which is in partial_update request) as the active.

*Proposed Solution:*

Introducing typedef option "appendRelationshipsOnPartialUpdate" which will hold list of relationship attribute names. While Atlas startup, it will resolve entity type definition and all the specified relationship attributes definitions will be updated with typedef property "isAppendOnPartialUpdate" only for that entity type. 


With this solution, we can keep all the previous relationship entries active along with latest relationship entry for the specific entity by providing typedef options as below.
"typeDefOptions": \{ "appendRelationshipsOnPartialUpdate": "[\"inputs\", \"outputs\"]"}

 

 



--
This message was sent by Atlassian Jira
(v8.3.4#803005)