Handle service not available errors in setup

When setting up workloads we create query/fts indexes. These steps will fail when the query/search service is not available on the cluster. This can be the case if a custom markov chain or the only-operation flag is being used. We should handle the service not available error and log a warning message that any operations using the relevant service will fail.

Start by locating the workload setup code that creates query and FTS indexes, then inspect how service-unavailable errors are currently handled. The work is done when setup logs a warning explaining that operations using the unavailable service will fail, while the relevant setup path handles the error for custom Markov chains and the only-operation flag.

Written by the indexing model from the issue text.