I’m sure that every Irreal reader is aware of the recent Internet outages caused by problems at AWS and Cloudflare. Most of those readers also know that things happen. Servers fail, cables break, backhoes wreak havoc, software runs up against implicit boundaries, and human beings make configuration errors. When those things happen, sites can lose their Internet connectivity. When they happen to services like Cloudflare or AWS, lots of sites can be knocked offline.
Cloudflare and Amazon are big companies with an outsize presence. That makes them an irresistible target for some people. Especially people who make their homes under bridges. An unfortunate number of these people are in the tech press which engaged in a feeding frenzy over unfortunate events that are rare but inevitable.
Sadly, even normally temperate folks can be drawn into the fray. Rik Huijzer, for instance, recommends that “ordinary” people not use Cloudflare because it’s a single point of failure. But what are the alternatives? At the end of the day you have to connect to the Internet and that connection point is just as apt to fail—more likely, actually—as Cloudflare or AWS. That’s true even if you self host: problems like this are almost always going to be at the network end.
I was inclined to dismiss the whole brouhaha with a shrug until Miguel Arroz injected a certain amount of rationality into the discussion and made me realize that reasonable people shouldn’t let this nonsense stand unchallenged. Arroz makes a great point: “Their downtimes being so visible is a consequence of their success.”