I think there are plenty of other examples of this fragility, such as free 3D model searching and services like that ADS (airplane call-sign) aggregator site or various weather-station front-ends.

And yes there's definitely a gray area where the distinction between indexing-only and data-hosting lies; definitely depends on what you regard as a "big" file. And if that's changed since last year.

But still. A #commonindex matters a lot & shouldn't be taken for granted when it's running smoothly.