"Under severe stress, a critical system in our infrastructure got overrun, triggered by a small error in our backend service communications." This was not due to a spike in external traffic or a specific event. The issue was instead caused by the increased number of servers in our datacenters.