The ultra-diffuse galaxy Dragonfly 44 is discussed in "Spectroscopic Confirmation of the Existence of Large, Diffuse Galaxies in the Coma Cluster" by van Dokkum et al (https://arxiv.org/abs/1504.03320). The preprint here is the source for all information about the galaxy in this answer.
The thing to remember about galaxies is that they generally contain a fairly stupendous number of stars. For example, we know that the Milky Way has on the order of 100 billion stars in it. We also know that Dragonfly 44 is roughly 100 times dimmer than a typical elliptical galaxy. According to the preprint, this figure actually refers to the total brightness of the galaxy (i.e. its absolute magnitude), not its brightness per unit area, so we don't need to take into account the relative size differences. Specifically, what the preprint says is that Dragonfly 44 is overall 5 magnitudes dimmer than a typical elliptical galaxy of similar size, which corresponds to a factor of 100 difference in total light intensity.
Since intensity is additive (2 average stars gives you twice the intensity as 1 average star), we can also very, very roughly say that Dragonfly 44 has 100 times fewer stars than the Milky Way (this is overlooking the fact that the Milky Way, a spiral galaxy, is not directly comparable to elliptical galaxies, due to being a different shape and also having a different type of stellar population, but neither of these differences should affect the results by more than a factor of ten).
This would mean that Dragonfly 44 is composed of somewhere around 1 billion stars. This is still a huge number of stars, much bigger than the largest objects we classify as globular clusters (which can be thought of as the next category down in size), so it's definitely still a galaxy by any reasonable standard.
It's true that the galaxy is almost entirely composed of dark matter, but galaxies in general are so massive that even a tiny fraction of their mass corresponds to an enormous number of stars, as you can see here.