We've previously been caught out by our functional tests running against the latest release of KubeVirt and not HEAD. For example when testing common-instancetypes during 4.13 where we only discovered bugs after a new release of KubeVirt had been cut and SSP had already been imported into HCO.

To avoid this it would be useful to add an upstream job deploying the operator onto a kubevirtci environment with KubeVirt deployed from source and running a limited set of functional tests against operands capable of running upstream without Openshift, at the moment these are common_instancetypes, datasources, tektonTasksOperand and tektonPipelinesOperand.