Does KafkaConnect use GCP's Workload Identity Federation?

[jymaeng95]

I want to use GCP Pub/Sub Source Connector using KafkaConnect, KafkaConnector CR.

GCP recommends using the workload identity federation to access resources in the cloud.

At this time, ServiceAccountToken should be mounted on the projected volume at the time the pod was created, but Strimzi KafkaConnect Spec could not determine the specification for mounting the projected volume and the specification for defining ServiceAccountName.

Is there any way I can use workload identity federation on KafkaConnect?

This is a link to the GCP guide I referenced.

https://cloud.google.com/iam/docs/workload-identity-federation-with-kubernetes#eks

Thanks.

[scholzj]

I think you might need to translate the Google doc to the Kubernetes level and share it here for people who have no experience with GCP's Workload Identity Federation - to explain what exactly is it you need / want. In general, Strimzi allows you to mount Config Maps or Secrets as volumes or environment variables. You can also - as with any Kafka Connect - use configuration providers.