[BUG] Shipwright Operator Creates Invalid TektonConfig

[adambkaplan]

Is there an existing issue for this?

• I have searched the existing issues

Kubernetes Version

Kubernetes v1.30 (provided by OpenShift v4.17)

OpenShift Pipelines Operator v1.16.1 (provides Tekton v0.62.z)

Shipwright Version

v0.13.0

Current Behavior

Shipwright Operator has logic to create a "default" TektonConfig object. With newer versions of the Tekton operator, a validating webhook now requires pruner settings to be added to the TektonConfig spec - one of spec.pruner.keep or spec.pruner.keep-since.

Due to a potential race condition with Operator Lifecycle Manager, the Shipwright operator can create this "invalid" TektonConfig before Tekton's validating webhook is up and ready. This leads to the Tekton operator failing to reconcile Shipwright's TektonConfig object, and thus preventing Tekton Pipelines from being deployed.

Expected Behavior

Shipwright Operator deploys a valid TektonConfig object, which leads to Tekton Pipelines being installed on the cluster.

Steps To Reproduce

With OpenShift 4.17:

1. Install Builds for OpenShift from OperatorHub (Red Hat's distribution of Shipwright)
2. Wait for the Builds for OpenShift and OpenShift Pipelines (Red Hat's version of Tekton) to be installed.
3. Check that Tekton and Shipwright's controllers have been correctly deployed.

Anything else?

Equivalent Red Hat issue tracker: https://issues.redhat.com/browse/BUILD-1199