On the one hand, the developers want bugs that are in the code that need to be fixed.
On the other hand, how is QA supposed to know whether it's code, data, scripts, or human error? QA knows that it doesn't work. QA reports that it doesn't work. What happens if it appears to be because the delayed job wasn't restarted, but it was in fact a bug in the initializer of the delayed job? QA isn't equipped to figure that out.
It's not QA's job to figure out the problem or solution. QA's job is to find the problem, and find the conditions that cause the problem, then communicate their findings. It's the job of the Product Manager to assign it to the appropriate person to fix it, be it a developer, DBA, systems administrator, or gopher.
So yes, you should raise a bug. It's not working, that is all you know. Whoever is doing the deployment should create a script to automatically restart delayed jobs when deploying. The fact that they don't, or that it isn't working, is the bug here. OTOH, if it's a problem because the job is in the process of restarting, the bug instead is that it isn't providing sensible feedback to the user ("Please Wait...").
Of course, if your boss says that you shouldn't be raising a bug, then don't. Your boss is the ultimate authority, not some random strangers on the internet.