A SaaS offering? Might surely be interesting for some.
There might be a few things you’d want to ensure a few things so the service can deliver value month after month…
- stand up a reasonably sized dev/ops team (as you wrote, this is a product; accordingly design & bringup is just 20% of the lifetime effort)
- an independent team
- let the team design a suitable stack (historically, that could become a fight against “not invented here” attitude)
- give that team some decision power for roadmap choices/priorities that concern their service
- or at least weigh their needs according to current - and projected - revenue
- give them a separate channel for raising security issues and obtaining fixes (different liability stance)
- scalability issues(*) would matter more since they hurt the business case
- the team’s documentation and support must be able to handle many small sites with their different, fractured patterns
If you put this teams’ needs behind the customers’ needs you’ll be forgetting that they’ll be one of the biggest users. they have a unique constraint regarding availability: the value proposition in the short run might be often just billing needs or similar.
but what if you look at the long-term success? Then it matters that it’s run consistently better than what makes those users look for SaaS.
This might mean that “we’ll have separate people on this” is an incomplete perspective.
You’ll have separate people that will have goals to reach that are operational and, as such, can collide with and be more important than product roadmap.
You need to be able to handle this, effectively a second pipeline that needs to stay filled. (one manager called Ops a “projektfeindliches Umfeld”). It will not be often that your main devs need to drop their work to ensure the SaaS can run smoothly - but if the time arises, they need to be ready to support their colleages at that time.
That’s the nature of offering a “service” 
I hope you’ll have a lot of luck with the new product, and to me it also sounds like it would have the potential to bring some more improvements (instead of novelties), so the frustration that has been expressed here might also be remediated.