CDN edge (hit rate + origin egress)
A CDN keeps copies of a customer's files near their users so pages load fast no matter where the visitor is. When a file is already cached nearby that's a "hit" -- served instantly. When it isn't, that's a "miss," and the request has to travel all the way back to the origin server, which is slower and costs you bandwidth.
The number that matters is the hit rate. At 85% hits, only one request in six ever bothers the origin. Let the hit rate collapse -- a cache purge, a sudden flood of rarely-seen files -- and origin load multiplies overnight along with your egress bill. Media and streaming customers obsess over this, because for them a slow tail is a lost viewer.
Detailed explanation
Hit rate is the lever: at 85% hit, only 15% of requests touch the origin. Misses cost origin egress + the round-trip latency; a hit-rate collapse (cold cache after a purge, or a long-tail content mix that doesn't fit cache) multiplies origin load and the egress bill simultaneously. Media/streaming tenants are judged on edge tail-latency, not average -- a good mean with a bad p99 still churns them.
A CDN sits in front of an origin -- usually an object-storage bucket or a load-balanced VM pool. Region count drives POP coverage: more regions = more edge locations = better hit locality. The cheapest way to cut a runaway egress bill is to raise the hit rate (longer TTLs, smarter purge), not to add bandwidth.