It's now nearly four years after "the great dimming of Betelgeuze" - and the matter has not yet been finally settled; as such this answer will necessarily include some personal judgement and weighting of the existing literature - and might be proven wrong as research progresses and things evolve.
Most evidence from modelling, radial velocity measurements and spectroscopy seems to suggest that it indeed was (quite a lot) ejected matter from the stellar atmosphere (into the direction of Earth) which cooled and in doing so obstructed the view from Earth onto the hot(ter) surface of Betelgeuze (images by E. Wheatly (STScl) and Montarges et al (2021) respectively):
Betelgeuze is known to be a semi-regular variable subject to radial puslations of about 400 days and a longer frequency one of a bit more than 2000 days. These variations can be understood and modelled. These models are e.g. presented in Montarges et al (Nature, 2021) or Macleod et al (ApJ, 2023), and also supported spectroscopically e.g. by Dupree et al (ApJ, 2020), Dupree et al (2022) or Alexeeva et al (Nature Communications, 2021). Prior to the dimming Betelgeuze showed an unusually long period of radial expansion, and subsequent spectroscopy is compatible with ejected matter cooling.
E.g. spatially resolved from Dupree et al (2020) show an asymmetry in the maximum emission at different wavelenth, supposedly indicative of a localized, unusually large outburst:
The radial motion is about anti-cyclic to the brightness of Betelgeuze. It has been unusually long been outward prior to the big dimming (Dupree et al (2022)):
Interesting in this context is IMHO also the paper by Neuhäuser et al (MNRAS, 2022) which study the colour of Betelgeuze based on historic records over the last few millenia; even when that is to be treated with caution and has big uncertainties, it hints at a recent redding, and thus also at Betelgeuze being in the beginning of its red giant phase rather than the end - thus making mass ejections as typical to these kind of red giants also a likely explanation.
Studies concerning chemical composition and modelling the pulsations (e.g. Saio et al (2023, in review)) are not really conclusive (e.g. see also this comment by Molnar et al on that preprint) as for the time till supernova the inner chemical composition is important, but only the atmospheric composition is accessible to spectroscopy and pulsation alone not compatible with spectroscopy. Thus even with optimistic assumptions in regard to a late stage of Betelgeuze, an outburst hypothesis is compatible with what is modelled and not an indicator of any other process.
Related sidenote: Notworthy is for all studies, that the mass of Betelgeuze is relatively hard to determine due to its rather big disc size (50mas) compared to its parallax motion (6mas) - that makes parallax measurements inaccurate. Thus luminosity is inaccurate due to distance inaccuracy, and thus also the mass determination. And basically every model requires mass as an important input parameter; this results in masses ranging from 12 to 20 solar masses being compatible with these observations, with the likely masses being somewhere in the middle or slightly above.
