When building pipelines using elyra in EMR, I have no choice but to use KUBERNETES SECRET.

[nanaones]

Is your feature request related to a problem? Please describe.

First of all, I'm using elyra very well. Thanks to the elyra team.

I'm envisioning MLOps with elyra and designing prototypes for ETL pipelines.

However, there is one disappointment.

I am using jupyterhub and Elyra on AWS EMR. Jupyterhub above EMR is authenticating using an IAM Role. In Airflow above EKS, I want to manage authentication via IRSA, but if elyra's auth_type is AWS_IAM_ROLES_FOR_SERVICE_ACCOUNTS, Jupyterhub above EMR will also unconditionally search AWS_ROLE_ARN or AWS_WEB_IDENTITY_TOKEN_FILE, so it throws an error.

*So I have no choice but to use KUBERNETES SECRET.

In EMR, I want to use elyra through IAM Role based authentication, and in EKS I want to be authenticated through IRSA.

Describe the solution you'd like A clear and concise description of what you want to happen.

When using the pipeline, the environment where Airflow is installed should be k8s, but the environment where elyra is installed may not be k8s.

It would be nice to be able to choose separately the cos authentication of the environment where elyra is installed and the authentication of the environment where the pipeline will run.

Describe alternatives you've considered A clear and concise description of any alternative solutions or features you've considered.

I think it would have been clearer if elyra had branched according to the environment in which it was run before classifying the if statement as an authentication method.

cos.py