Parameterize the number of retries for apply_workloads

For certain workloads, which involve the deployment of operators via OLM, we need some time for that to be applied properly. Those operators sometimes define custom resources, which cannot be applied until the operator is successfully deployed.
That is the case of PerformanceProfile, which is failing on https://jenkins.akraino.org/job/kni-blueprint-pae-verify-deploy-baremetal/29/console . This can be solved just by iterating, but for our apply workloads, we just retry 2 times.
We need to be able to parameterize the number of retries of knictl apply_workloads (it's fine to default to 2). If we add this parameter, then we can increase the number of retries from our ci script.