Most of the time these status pages sit on, and store data in, the same infrastructure as the things it's supposed to be reporting on.

Do they really? Amazon did that one time but that was stupid of them. Most software companies use an external service (like statuspage.io) to host their status page.

In my experience, changing the status pages at many companies involves a C-level decision, since that status page change serves as an admission of service outages. And service outages can trigger availability clauses in contracts, costing the company significant amounts of money.

The mental acrobatics to justify not changing the status are often astounding: "One customer can load a partial page from Timbuktu, so we're not down."

Probably the case. I've seen the same thing with Heroku. I've report an issue, then half an hour later it gets updated. Could also be a bad cache that takes a while to invalidate.

HN Could make a killing selling a saas product that monitors if sites are up. Honestly it's more reliable at this point.

Ostensibly, that already exists: https://www.statuspage.io/

I believe OP's point was that aggregating crowd-sourced reporting of outages is surprisingly accurate and prompt, not that there aren't any other tools for reporting outages.

We've actually created just that. https://hund.io We're a status page provider with built-in monitoring so if a service goes down the page will be updated.