One important caveat to all this is that the gravitational redshift for a galaxy or galaxy cluster is an incredibly tiny effect. For example, the mean gravitational redshift for massive galaxy clusters has been measured (e.g., link below), and has a value of ~ 10 or 15 km/s. This can be compared with the typical motions of galaxies within a cluster (dispersions of ~ 1000 km/s) and the cosmological redshift of the clusters (for the sample in the paper I linked to, it's equivalent to an average Doppler shift of ~ 50,000 km/s).

In order to make that measurement, the authors had to combine redshift measurements for ~ 20,000 galaxies from ~ 2,500 different clusters; as they note, it's impossible to detect the effect for just a single cluster: "... even assuming that every galaxy in a cluster could be spectroscopically measured, there are simply too few galaxies to allow the statistical detection of a non-zero gravitational redshift."

So while it's technically true that adding dark matter to a galaxy would increase its gravitational redshift and thus its total observed redshift, this effect is immeasurably small.

Mpetha et al. (2021), "Gravitational redshifting of galaxies in the SPIDERS cluster catalogue" https://arxiv.org/abs/2102.11156