I don't think we can ever prove one way or the other if we're in a simulation. I am not really sure it should actually matter to us though. But in answer to a couple of the points you make.
the bugs would be obvious malfunctions in the code. But would the program and those in the program realize they are bugs? A sentient NPC in GTA, for example, would they realize the car that just glitched through the world is not normal behavior? Perhaps the bug also affects their understanding of their world too.
If there are bugs, we'd likely not know it. Because everything about the world and universe around us is normal, because it is how it is "warts and all". But specifically we have some odd things like some of the effects attributed to quantum theories. Perhaps they could be considered bugs.
our simulation system is rebooted on a normal basis but we never see it which reduces the bugs observed. Perhaps the planet operates on a docker-like platform, and when everyone in the section is asleep, the system is rebooted unbeknownst to the users residing there. Or reboots are not observed by us and we have no perception of “lost time”.
Why would you see it? If this is a simulation, and the entire system's state is frozen, stored and the system shutdown for 1000 years. Then restarted, for us no time would have passed, and we'd be unaware of the "shutdown".