Our IaaS is excelent, but nothing is perfect: 3h downtime across all servers
Our Infrastructure as a Service partner, Nexeon, has served us for almost 2 years without a single failure. Of course they put a lot of effort in that: Kevlar in the walls, double power redundancy on no-breaks plus a set of generators that can keep the facility online without power for days, 16 optics fiber entry points, 3 traffic lanes of data-link, spare hardware for anything and more (continue reading).
But, problems happens and take time for noticing and get fixed. After 50 minutes of down connectivity for all the IPv6 links, across all the servers, we noticed the problem and got in touch to Nexeon. It took 2h to solve the problem in the IX. Not a single node with IPv4 was affected.
The timeframe of the solution is acceptable but promoted losses to our users. We will provide a full reimbursement to our user affected by this issue and will claim it with our partner.
Despite that, the high cost charged by Nexeon due its premium service made worth every penny. We are confident that they will offer even better quality metrics than the one observed in this previous 2 years: 3hours of downtime for IPv6 (that is pretty good already)