You list three conditions that your problem must meet, but then say that number 2 is an "I would like." So I'm not sure if we must meet number 2.
Systematic weakness in the buoyancy providing 'balloons'.
First, I am working with an assumption of the most practical Venusian cloud city. A large, lightweight colony is constructed and pressurized to normal atmospheric pressure, and then held aloft by balloons. The balloons have a lift gas (doesn't really matter what gas, as long as it is lighter than CO$_2$) and are pressurized to atmospheric pressure at that altitude, roughly 0.5 atm at 55 km altitude.
In this case, there is no 'engine' because the whole colony floats instead of having a propulsion system. There would probably be some stabilizing propellers needed to prevent rotation, but no ability to move around in the skies. There would be 'engines' to provide power, the nature of which you haven't really specified. There might be either a fossil fuel or nuclear fission generator for nighttime electric generation when solar isn't available. More likely, you could assume there is a great leap forward in batteries or hydrogen fuel cells, and those can be used in conjunction with solar.
In either case, the electric power system won't be causing the disaster, it will be one (or more) of the buoyancy balloons. They were perforated by some sort of impact event, and the perforations are right along a mechanical high stress area, so they are vulnerable to sudden catastrophic failure.
The great advantage of the lift gas being at the same pressure as the Venusian atmosphere, is there is no sudden decompression event. The CO$_2$ atmosphere and the lift gas will slowly mix as the balloon loses buoyancy.
I believe this meets your criteria (minus mentions of the engine, which, as I explained, I don't think this station will have).
The perforations are small and hard to detect. They wouldn't be a problem if it weren't for their orientation along one of the highest stress zones of the balloon.
This is a gas leak, and although it probably wouldn't be heard, you could have it found in a routine inspection of the buoyancy balloon.
This problem will inevitably cause the colony to descend.It won't happen right away, even if the balloon rips, but action must be taken BEFORE the perforations become a major leak.
Regarding point 2, your description of a whistling sound coming from an air vent in the engine room does not describe an actual problem. I have experience with a nuclear fission reactor, a 600# steam plant, gas turbines, and diesel engines. I have actually heard leak in both the steam plant (downstream of a regulation valve, thank god) and diesel plant (jacket water piping, in case you are interested). Both leaks are not 'a very faint whistling noise.' They were an immediate existential crisis for anyone in the same space. Engines run at high temps and high pressures. Leaks tend to be fiascos. I don't think you can get number 2.