Netflix and verizon do indeed have pops in the same facilities. There was a recent post by a verizon suggested that all the congestion between netflix and them was due to a 100% saturated interconnect in LA to L3. That of course means that *all* verizon traffic going through that node is badly impacted. Verizon and netflix are currently working out a deal to peer directly. This will likely entail several hundred 10g ethernet interconnects spread out across numerous locations. But until then verizon is happily screwing over their customers just for a leg up in negotiations with netflix.

By far the better solution is to figure out why that one specific server cant be offlined. Its far safer regardless of the tests and validations to work on a server thats not supposed to be running vs one that is. It obviously takes alot of work, but for all your critical/important services they should be running in some sort of HA scenerio. If you cant take a 5 minute outage just after normal business hours, you absolutely cannot take a failure in the service due to any sort of hardware failure(which will happen) This is coming from years of experience in a Software as a Service company/