Others should help, but I can’t think of big forum regression reports on 2.0.4.18 canary though I’m always wondering whether canary has enough (brave) users running it in non-critical production-similar manners. Maybe some more users will try it out on the Experimental channel if that serves as a pre-Beta test again.
Usage reports do show the version in use, but I’m not sure how good a view that gives of canary’s usage. Some of the proving of things like network error handling of parallel uploads may need time and exposure.
Commits after 2.0.4.18 (out a month) look heavy on code tidying, plus a few enhancements and fixes that maybe could be nominated for inclusion by committers (and tested somehow?). I’d nominate issue 3787, except nobody seems willing/able to commit it. This concerns me because it’s a data corruption situation.
#3790 is a privacy regression in DB bug reports, but those are manual and they seem not much used but maybe should be if we hope to ever fix the tough-to-track-down backup corruptions that sometimes occur.
To further address the original question, Beta releases are fairly rare. Changes go into canary, break stuff sometimes, and eventually things stabilize, and Beta (and maybe Experimental) goes on to a wider group:
v2.0.2.1-2.0.2.1_beta_2017-08-01
v2.0.3.3-2.0.3.3_beta_2018-04-02
v2.0.4.5-2.0.4.5_beta_2018-11-28