Expected Behavior

When deploying Tekton Pipelines, I expect to be able to change resource names to match our conventions.

Actual Behavior

• The webhook Deployment and Service must be called tekton-pipelines-webhook
• The logging and observability ConfigMaps must be called config-logging and config-observability, despite being configurable in the controller’s Deployment
• The configuration for bucket and PVC artifacts must be called config-artifact-bucket and config-artifact-pvc

Steps to Reproduce the Problem

1. Change any of the above resource names in the Tekton Pipelines deployment YAML bundle
2. Apply them
3. Witness the CrashLoopBackOffs

Additional Info

We have a policy of not blindly applying .yml distributions to our clusters; instead, we go through them to make sure we understand what we’re deploying and how the services that comprise a system fit together.

There’s almost always things in the distributed resources that we like to clean up; for Tekton, I thought it was redundant to name resources tekton-pipelines-… when we were creating a whole tekton-pipelines namespace for them; and likewise redundant to start all ConfigMap names with config-….