In terms of multivariate analysis — sure, it’s true to say that every observational study ever has potential confounders that they can’t explain away. Which is why it’s problematic to go from a single observational study, especially one which goes against quite a bit of established research as this one does, and say that eating chocolate is healthy. It’s not so much the analysis that I take issue with, it’s the news articles (and to a certain extent the researchers) arguing that this implies that chocolate is healthy when there are a huge number of potential factors that they did not account for.
In terms of how healthy chocolate is; there is a common misconception that dark chocolate, because it is bitter, is healthy. Whilst there may indeed be some healthy form of dark chocolate, pretty much any dark chocolate that you can buy commercially is >15% sugar and >50% fat, making it hard to justify calling it healthy (all of those links were to so-called ‘healthy’ dark chocolates, supermarket brands are far worse). Dark chocolate is the raisin muffin of sugary snacks: taking a little bit of unhealthy out doesn’t suddenly make the rest good for you. The flavanoids that may be healthy don’t really make up for the fact that you are eating something that is basically delicious fatty cocoa butter with some added sugar.