I would still prefer the job being scheduled in the agent. This avoids putting all your eggs in one basket by counting on the one scheduler to be maintained correctly. (I don't count on our SAN to not be down for 3 days. I don't count on our enterprise scheduler not stopping because it thinks the license expired.) There is a good solution that does most maintenance available already (). Ola's solution provides log files with details. It also has the option to clean old log files or backups. It should normal run without issue.

If there is no DBA to monitor the maintenance, then perhaps a few alerts to a non-DBA so that they let management know that they need to find a DBA to fix it.

If somebody needs a simple button on the desktop, then perhaps the PSH can get the maintenance history, check file space, start jobs, copy specific backup files, etc.

I have been thinking about using PSH to start jobs because we have been using several instances on some beefy hardware. It would be nice to be able to run related jobs (e.g., backup) in order across instances with one start. (I manually schedule them to not overlap now.)

Even if a PSH script is complex enough to replace Ola's solution, I would still want to run it in the agent. Perhaps as a PSH job step. Perhaps it could be instance aware to allow redundancy or complex scheduling? (I was thinking to create a job step using PSH to detect and wait if the same job is running on another instance of the same server. That would be easier than replacing Ola's solution.)