Cache hit ratio

Cache hit ratio is the share of requests a cache answers itself instead of forwarding to the origin: hits ÷ (hits + misses). It's the single most telling CDN metric because it's three metrics in one — speed (a miss is 3.5× slower at the median, per our measurement), cost (misses consume origin compute and egress) and resilience (a warm cache absorbs spikes your origin can't).

Request vs. byte hit ratio

Request hit ratio counts requests; byte hit ratio counts transferred bytes. A site can show 95% request hit ratio while large, rarely-cached downloads push the byte hit ratio far lower. For visitor experience watch the request ratio; for egress bills, the byte ratio.

What "good" looks like

• Static assets: 95%+ is achievable and expected.
• HTML: depends entirely on whether you cache it at all (many CDNs don't by default) — cached HTML above 90% is realistic with sane TTLs and re-warming.
• The aggregate number flatters you: it's dominated by your hottest URLs at your busiest PoPs. The long tail of pages, at the far edge, is where the misses live — we measured two orders of magnitude difference between locations.

The levers that raise it

In rough order of impact: make HTML cacheable at the edge (Cache-Control / s-maxage, CDN cache rules), clean up your cache key (strip tracking parameters), avoid hit-ratio-shattering Vary values, purge narrowly instead of purging everything, enable tiered caching, and warm the URLs organic traffic doesn't keep resident. The full playbook with measurement instructions per CDN is in our cache hit ratio guide.