Dangerous is a word that doesn't actively justify trade-offs. With SaaS offerings, you're trading opex costs to save on capex costs, which can mean a really big

Dangerous is a word that doesn't actively justify trade-offs. With SaaS offerings, you're trading opex costs to save on capex costs, which can mean a really big difference in terms of opportunity costs. You also are paying for the expertise that you'd need to develop internally otherwise. These aren't simple black and white choices. Operating at FB level scale means that saving 2% on your cost of revenue is billions of dollars. If this is the only "Big" outage FB has for the next 2 years, and it's saved them billions of dollars, isn't the fairly small risk of a single point of failure worth it? You have to accept some risk, and it's often hard to compare risks on the business side vs risks on the technical side.