[OP] the decay of reliable infrastructure
blog.dijit.sh
blog.dijit.sh
Then I move to AWS and I get a call from a crack salesperson who says that "people who leave us for AWS always come back" and you know... I haven't gone back.
And that was a reputable hosting company. Back around 2001 I was working with an entrepreneural minded MBA and he would find incredible deal for dedicated servers from companies that would suddenly go out of business or have their suppliers go out of business -- we got fast at restoring from off-site backups so it was ok...
If you need reliabilty, use a virtual machine on an IBM 390.
Typically I'm talking on a relatively big scale and I should have mentioned that in that poor excuse of a blog post.
We're a pretty big company, we have our own datacenters our own hardware. And usually when you work with a half-decent server provider you talk directly with a person who has physical access to the machine and the ability to change things like that.
Still- that doesn't discount your experience and if you're at VPS/Startup/Small infra level then cloud is a fine solution to be sure.
My experience of the cloud (amazons in particular) is that instances just "go away" for 8 hours and you cannot contact support, the little service alert says the availaibilty zone is green but you have a little dialog saying "you may experience some issues with your instance".
Which happened to be 8 times in one month.. And I'm sure that no client is big enough for amazon to start giving a solidary fuck about customer service in that regard.