Given that caffeine can be readily extracted from a variety of abundant and exceedingly cheap natural sources, I doubt it would be economical to synthesize. Various decaffeinated products are popular, and the caffeine removed during the decaffeination process is normally recovered, purified, and re-sold for use in various products, which makes the synthetic route even less economically attractive. As far as I'm aware, the vast majority of caffeine sold commercially is extracted from plant sources. You can do your own research to determine the exact statistics, but since your concern relates to caffeine extracted from natural sources that might also contain harmala alkaloids, I'll address that case specifically.
There are a number of different methods of caffeine extraction (see the Wikipedia page on decaffeination), but all use solvents and/or adsorbents at some stage in the separation and purification processes. From the perspective of chemistry, then, the question is how similar are the solubilities of caffeine and harmala alkaloids, and what percentage by mass do they constitute of the natural caffeine sources used? I was able to find one article that quantified the chemical constituents of coffee beans. If you examine the data table therein, you'll see that the main constituents of those beans are carbohydrates, lipids, amino acids, and chlorogenic acids, not one of which is a category that applies to the harmala alkaloids. Lipids, carbohydrates, and amino acids are generally the main constituents of virtually all plants, so I think it's probably safe to extend those statistics to other natural sources besides coffee beans. Purely as a matter of probability, it's reasonable to assume the impurities are likely to consist mostly of substances from those categories. With respect to solubility, the harmala alkaloids are mostly somewhat heavier nitrogenous bases which are less polar than the caffeine molecule (at least in their neutral, "free base" form). Given that, I'd expect that methods using water extraction, which appear to be popular, would separate caffeine from harmala alkaloids fairly efficiently (I found one paper indicating that harmala alkaloids are water insoluble, while caffeine is anywhere from moderately to highly soluble depending on temperature).
An impurity of 1% is already quite small, and that 1% comprises a variety of compounds, both organic and inorganic (the overwhelming bulk of which are likely not harmala alkaloids, given the natural composition of caffeine-containing plant sources I described above). Based on the above information, I'd suspect that only a tiny fraction of those impurities, if any, would be likely to consist of harmala alkaloids at a detectable level.