Does 99.995% [1] of hackernews sound reasonable enough to you?
The reality is a lot of systems (especially simple ones) run perfectly fine on a single server with next to no downtime and all the additional redundancies we introduce also add additional points of failures and without the scale that makes these necessary you might actually end up reducing your availability.
The reality is a lot of systems (especially simple ones) run perfectly fine on a single server with next to no downtime and all the additional redundancies we introduce also add additional points of failures and without the scale that makes these necessary you might actually end up reducing your availability.
[1] https://hn.hund.io/