In December of 2020, we got very nervous.
Let's go a little back in time. We ran our very first infrastructure on a small German data center.<p>They offered a really nice UI, were GDPR compliant, and also much cheaper than the big international competitors in comparison.<p>What we were not aware of, was that they provided a maximum of only 10 servers.<p>This meant that there was no way they could scale with us long term, as our team is located in different countries and utilize hundreds of nodes. It would have not even sufficed for our launch alone.<p>Our dream requirements were:
initially 80+ servers in one German datacenter
each server having >50GB RAM and >10 vCPUs
>20% automatable growth per month
low price per node
easy monitoring
managed load balancer (HAProxy)
simple reimaging
dedicated NVMe SSD’s for our ETDC cluster
GDPR compliant<p>In the long run, we want to scale to hundreds of servers per data center. We strive to have multiple data centers in the USA, Asia, Europe, Russia, and Latin America with ideally, as few providers as possible.<p>After having 20+ calls with hosting providers, it seemed that all of them would either be more expensive or did not provide what we required at scale.<p>We realized that it was unrealistic to secure and maintain our own hardware and keep our launch schedule on track at the same time.<p>Although we found multiple providers such as Contabo, SysEleven, and PlusServer, that could get the job done, we ultimately decided to collaborate with Plusserver to build our own datacenter on their new GAIA-X infrastructure.<p>We believe that trusting your infrastructure to be able to scale smoothly with you is the key requirement to be able to focus on your product.<p>We are more than thrilled to be working with the guys from PlusServer and whether or not this was a success, you can be the judge of, after testing Codesphere. :)<p>What are your experiences with infrastructure decisions? What can we do better?