1. Proposed title of this feature request

Clearer Error Messaging for Registry-related ImagePullBackoff

2. What is the nature and description of the request?

There are numerous reasons why Kubernetes can't pull a container image and run it in a pod. Troubleshooting this situation is particularly challenging when multiple controllers/components are involved. For example - OpenShift Pipelines failing to run a task because Kubernetes can't pull a referenced image in the task definition.

Currently there is no easy mechanism to surface the root cause of these image pull failures. While not all error situations can report a meaningful message, HTTP error codes from container registries can translate to a more meaningful error message, such as:

• Unauthorized/image needs pull secret (401)
• Access to image forbidden (403)
• Image does not exist (404)
• Rate limit exceeded (429)

If the error messaging is improved in the pod/container status, this information can then be bubbled out to other components that infer state from pod/container status.

3. Why does the customer need this? (List the business requirements here)

Surfacing registry-related issues for ImagePullBackoff helps end-users identify issues that prevent workloads from running. This ultimately improves the experience of the entire OpenShift/Kubernetes ecosystem.

4. List any affected packages or components.

OpenShift
Builds for OpenShift
OpenShift Pipelines

... anything that forwards pod status to a Kubernetes Custom Resource