[contingencies]

Yes, it's nice to see people thinking of this.

What about IPv6? Non-standard protocols? Layer 2 connectivity requirements? Same across multiple data centers where latency and packet loss are potentially higher than negligible and volatile? Link disruption handling? Failover protocol? Service ordering? Distinction between startup and execution time dependencies? Who to call/notify when something breaks? How to do so? How to escalate? Integration of high security requirements such as multi-party authentication and signoff using crypto and/or multi-factor (phone, sms, crypto-devices, etc.)?

There's so much to do here, this post is barely scratching the surface.