Anyone know what Twitter's infrastructure deployment process was like? I assume that if anything will go wrong, it'll start there; the software itself shouldn'

Anyone know what Twitter's infrastructure deployment process was like? I assume that if anything will go wrong, it'll start there; the software itself shouldn't need to be handheld so much. If it's all Kubernetes and what-not, and it still too complex for mere-mortals to grok then all companies should be worried.