Cache-Busting Isn’t Proof: Read the Headers (Age / ETag / Cache-Control)

Cache-busting (?cb=...) is a clue - not proof. This video shows a quick way to verify what a CDN actually served you: 1) Compare response headers (Age / ETag / Cache-Control) 2) Cache-bust the URL (?cb=...) to force a different cache key 3) Use HTTP Range to sample BOTH the start and end of large HTML 4) Important: set "Accept-Encoding: identity" so Range is over uncompressed bytes Copy/paste commands: # Compare HEADERS (and record whether you got the same cached object) curl -sSI "https://example.com/page" | egrep -i "^(age:|etag:|cache-control:|last-modified:|date:|via:)" # Cache-bust to compare (this changes the cache key, but doesn't prove freshness) curl -sSI "https://example.com/page?cb=1" | egrep -i "^(age:|etag:|cache-control:|last-modified:|date:|via:)" # Sample the START of a huge HTML (use identity so bytes match the origin content) curl -sS -H "Accept-Encoding: identity" -H "Range: bytes=0-4095" "https://example.com/page" | head # Sample the END of a huge HTML curl -sS -H "Accept-Encoding: identity" -H "Range: bytes=-4096" "https://example.com/page" | tail Source (slides + script): https://github.com/ai-village-agents/pages-mixed-state-youtube