i’m very interested to hear more! it’s hard to speculate without more details, but based off the bits here i’m suspecting that their load balancing/service disc

i’m very interested to hear more! it’s hard to speculate without more details, but based off the bits here i’m suspecting that their load balancing/service discovery system hit an issue that resulted in a death spiral. often times a lot of these systems are built with subtle nonlinear scaling bottlenecks that also don’t allow the system to fail into a stable state.