The maturity of HTTP/2 is not a causative factor here. They removed a previously-unaware limit on the number of concurrent backend requests, which overflowed th

The maturity of HTTP/2 is not a causative factor here. They removed a previously-unaware limit on the number of concurrent backend requests, which overflowed their backends. They could have experienced this same outage by simply removing that limit without enabling HTTP/2, and then hitting a peak demand period that was sufficient to cause the outage. Yes, HTTP/2 changes traffic patterns, but the issue could easily have occurred with HTTP/1 as well.