Enzyme-based sensors for taking real-time measurements of growth substrates do exist. In brief: an enzyme degrades the substrate and thus changes the local redox state which is detected electronically (example glucose sensor). There's no guarantee that the substrates that you want to measure can be measured using such sensors. Also I think those sensors are fairly expensive and are tailored to specific conditions (ie it's not as simple as dipping the sensor into your culture medium). Practically speaking I think systems using these sensors are limited to monitoring just a small number of substrates.
LCMS or GCMS is the standard method for measuring concentrations of substrates. LC or GC (without the MS) may be possible, but probably not since there are are many components to culture media and so lots of overlapping peaks. Using this method is labor intensive (lots of samples), likely to be low resolution (temporally), and the equipment is pretty pricey, but if you have access to an LCMS and are willing to put the time in to become an LCMS expert then this is a good option.
Culture based methods using minimal media supplemented with substrates is an option, these experiments may achieve the same thing as real-time monitoring of substrate concentrations. If you have access to a plate reader this will be so much easier. The speed at which different substrates are consumed, and the order in which they are consumed can be inferred by measuring the growth rate of the bacteria and the final population size/dry weight of the culture.
I mention this culture-based method because, although it's not exactly what your question was about, it's much cheaper than the other options and much more established (and in my opinion more fun: you get a great feel for how the bacteria grow and what media they like, and the graphs look great).