Ask HN: Can “as a service” obviate Disaster Recovery for its failure modes?

Take DBaaS for example. A failing DBaaS categorically cannot, in my opinion, be the source of D/R for itself. I should be able to backup my data outside the wire so to speak if I want provide Disaster Recovery for my own clients. Could be NoSQL. Could be SQL. Could be BeyondSQL.

I'm interested in a counterpoint of my position that D/R is categorically "outside" the capability of a single "as a service". It's not that the service can't provide for minor to medium failure modes. It's that I may need to rollback to a data snapshot from last week on a completely different DBaaS provider when the DBaaS provider goes offline for longer than is tolerable.

SLAs are great, but do they pay me back what I lost in brand trust?