Details
-
Bug
-
Status: Resolved
-
Minor
-
Resolution: Fixed
-
1.12.1
-
None
Description
During Elasticsearch full cluster restart, PutElasticseachRecord routed incoming FlowFiles to its 'failure' relationship. In contrast, PutElasticsearchHttp / PutElasticsearchHttpRecord routed incoming FlowFiles to 'retry'.
While PutElasticsearchHttp processors determine if a request can be retried by checking if HTTP status code is 5XX, PutElasticsearchRecord and corresponding ElasticsearchError only check the name of thrown exception.
To make PutElasticsearchRecord more resilient against such situation, ElasticsearchError should treat ResponseException as recoverable where HTTP status code is 503 Service Unavailable.
Attachments
Issue Links
- links to