Closed anand3493 closed 2 years ago
Verified the fix in 5.2.2 release. working fine. Thank you for fixing the issue @cosmo0920
Wow, thanks for the conformation.
@cosmo0920 In one of the environment am getting the below error for the same packages:
Can u please check on this and let me know. As the stack trace suggests, it starting from fluent-plugin-elasticsearch only
This could be caused from non-existent [default]
profile on ~/.aws/credentials
or /root/.aws/config
.
@cosmo0920 I have not configured assume_role_arn for the fluentd as the destination elasticsearch is open to public. The stack trace suggests it looks for Elasticsearch major version and then trying to hit some aws api to get the same. However I have even tried to supply : verify_es_version_at_startup false default_elasticsearch_version 7 - these 2 fields so that it doesn't require to do so. But still am getting the error
I'm not sure why you are using almost obsoleted plugin that is fluent-plugin-aws-elasticsearch-service. We recommend to use fluent-plugin-opensearch. That plugin can handle Fluentd logs ingestion into the recent AWS OpenSearch service.
(check apply)
Problem
Getting error with fluentd - aws elasticsearch setup [error]: #0 unexpected error error_class=NameError error="uninitialized constant Fluent::Plugin::ElasticsearchOutput::Elasticsearch::UnsupportedProductError"
Fluentd Pods are running. But logs are not flowing. FLuentd Logs reported error with stack trace:
...
Steps to replicate
Fluentd daemonset with aws elasticsearch plugin is configured to push logs
Expected Behavior or What you need to ask
Fluentd to push logs to ES ...
Using Fluentd and ES plugin versions