Operator versions that were available today in the Red Hat operators catalog can be gone tomorrow. This happens when a new version of an operator gets released that is not referencing the previous version either by use of skips or replaces which causes the previously released version to be truncated from the catalog.

This is happening due to the way the ART pipeline processes new operator released. Operator released via CPaaS don't suffer from this issue by default but they can if misconfigured by the operator author.

Advanced Cluster Management for Kubernetes
Multicluster Engine
Topology Aware Lifecycle Manager
Local Storage Operator
OpenShift Data Foundation Operator
OpenShift GitOps Operator
Performance Addon Operator
SR-IOV Network Operator
TALM
Openshift-logging
mcg-operator
ocs-operator
odf-csi-addons-operator
ptp-operator
 

When a new version of any of the above operators (or any ART-managed operator) gets released the previously released versions in the channel disappear.

1. Install $operator on a cluster in version A
2. Wait until version B is released
3. re-attempt to install $operator in version A

Version A is not found in the catalog by OLM

Version is part of the catalogue and remains installable.

This has a large impact on customers who are using staging environments to test operator-based products before promoting a new version to production. Once a staged version is approved it may not be installable in product because ART has released a new version into the catalog, truncating the old version in the process