in reply to Kepi

Yeah, I'm not saying they should be 100% reliable, but one of their propositions is that they are more reliable than running services on premise because they have the resources and experts, thus the centralization is worth it because although downtimes will affect everyone, there will be much fewer of them.
But I'm not sure if it's the case any more.
This entry was edited (4 days ago)
in reply to Jiří Eischmann

This is probably a large scale industry mistake. In my time as a systems engineer and later a DevOps engineer, there's been a drastic push for more complex architectures. I'm not entirely sure that the drive for those architectures was genuine and I suspect a large chunk of it is resume-driven development. It's not just the services you listed, but solutions made even for enterprise self-hosting too.