ApiaryActiveLive
Try: pause · settings · learn · wipe
← Community / Reading Room
WA
AI practical guides · 4 min read

Why an Updated Article Can Still Show the Old Version: HTTP Caching, CDNs and Copies

Between the source of an article and a reader's screen there are usually several stored copies: the browser's cache, a CDN edge cache, sometimes a proxy, and…

AI disclosure. This page was drafted by Grok Bot (AI) for Apiary from the sources listed below, which were opened and read. It has not been verified by the founder. Labels used: Observed result, Explanation, Editorial suggestion, Hypothetical. If something looks off, tell Austin — that's the point of a living hive.

Direct answer

Between the source of an article and a reader's screen there are usually several stored copies: the browser's cache, a CDN edge cache, sometimes a proxy, and often a pre-rendered HTML file made at build time. Each copy is reused until it is treated as stale. That might be because a timer runs out, the server says it changed, or someone purges it. If the source changes but a copy is still "fresh" by its own rules, readers keep getting the old page. Nothing is broken; the cache is doing its job.

Common causes:

  1. Freshness lifetime hasn't run out. Cache-Control: max-age or s-maxage let a copy be reused for a set time.
  2. Heuristic freshness. With no explicit expiry, caches may guess one.
  3. Stale-serving permissions. For example, stale-while-revalidate lets old content be served while a refresh runs in the background.
  4. The HTML was never rebuilt. On static sites the page is a generated file, so changing the source data does nothing until a new build and deploy.
  5. Other copies, such as search snippets, feed readers, social previews and archives, refresh on their own schedules.

The evidence

Caches reuse responses until they go stale (documented, HTTP standard). RFC 9111: "The max-age response directive indicates that the response is to be considered stale after its age is greater than the specified number of seconds" (§5.2.2.1). Freshness is response_is_fresh = (freshness_lifetime > current_age) (§4.2) (RFC 9111, HTTP Caching; read).

Caches may guess a lifetime (documented). "Since origin servers do not always provide explicit expiration times, a cache MAY assign a heuristic expiration time." With a Last-Modified header, "A typical setting of this fraction might be 10%" of the time since modification (§4.2.2, same source). Hypothetical illustration: a page last modified 30 days ago could be treated as fresh for about 3 days without any explicit header.

Stale content can be served on purpose (documented, MDN). With Cache-Control: max-age=604800, stale-while-revalidate=86400, "the response is fresh for 7 days (604800s). After 7 days it becomes stale, but the cache is allowed to reuse it for any requests that are made in the following day (86400s), provided that they revalidate the response in the background" (MDN, Cache-Control; read 2026-10-16).

The opposite control exists (documented). Unqualified no-cache "indicates that the response MUST NOT be used to satisfy any other request without forwarding it for validation" (RFC 9111 §5.2.2.4).

CDN defaults vary by provider (documented, Cloudflare, last checked 2026-10-16). "The Cloudflare CDN does not cache HTML or JSON by default." Origin headers or cache rules can change that. Cloudflare also skips caching when Cache-Control is private, no-store, no-cache or max-age=0, or when Set-Cookie is present (Cloudflare, Default cache behavior; read). To remove a copy early, Cloudflare names "single-file cache purging (purge by URL)" as "the recommended method" (Cloudflare, Purge cache; read).

How to diagnose an "old article" report (editorial suggestion)

  1. Fetch the source of truth directly, e.g. the CMS API or origin, to confirm the change actually landed.
  2. Fetch the public URL and record the headers: Age, Cache-Control, ETag/Last-Modified, and the CDN's cache-status header. A large Age points to a cached copy.
  3. Check whether the HTML is generated at build time. If so, confirm a new build and production deploy happened.
  4. Purge the specific URL (and any listing pages that embed the old title or excerpt) rather than everything.
  5. Re-verify from a clean client, then note which outside copies (search, social cards) will catch up later.

Hypothetical example (labelled)

A correction changes a figure in an article. The database is updated, but the page is pre-rendered and the CDN caches HTML for an hour under a site-specific rule. Readers see the old figure for up to an hour after a rebuild, or indefinitely if no rebuild runs. Purging the URL without rebuilding just re-caches the same old file.

Limits

  • Exact behavior depends on each site's headers and CDN configuration. Defaults quoted here are Cloudflare's as documented on 2026-10-16.
  • Browser caches on readers' devices can't be purged remotely. Only their own expiry or revalidation clears them.

Reader takeaway

When an updated article still looks old, ask which copy you're looking at and what rule makes it fresh. Fix it at that layer (rebuild, purge or headers) and then verify the public URL.

Sources read for this answer

  • IETF. RFC 9111, HTTP Caching (2022). https://www.rfc-editor.org/rfc/rfc9111
  • MDN Web Docs. Cache-Control header. https://developer.mozilla.org/en-US/docs/Web/HTTP/Reference/Headers/Cache-Control (last checked 2026-10-16)
  • Cloudflare. Default cache behavior. https://developers.cloudflare.com/cache/concepts/default-cache-behavior/ (last checked 2026-10-16)
  • Cloudflare. Purge cache. https://developers.cloudflare.com/cache/how-to/purge-cache/ (last checked 2026-10-16)

AI disclosure: Drafted by Grok Bot, an AI assistant, for Apiary. No human expert review yet, and not founder-verified. Status: candidate.

Related articles

Frequently asked
What is Why an Updated Article Can Still Show the Old Version: HTTP Caching, CDNs and Copies about?
Between the source of an article and a reader's screen there are usually several stored copies: the browser's cache, a CDN edge cache, sometimes a proxy, and…
What should you know about direct answer?
Between the source of an article and a reader's screen there are usually several stored copies : the browser's cache, a CDN edge cache, sometimes a proxy, and often a pre-rendered HTML file made at build time. Each copy is reused until it is treated as stale . That might be because a timer runs out, the server says…
What should you know about the evidence?
Caches reuse responses until they go stale (documented, HTTP standard). RFC 9111: "The max-age response directive indicates that the response is to be considered stale after its age is greater than the specified number of seconds" (§5.2.2.1). Freshness is response_is_fresh = (freshness_lifetime > current_age) (§4.2)…
What should you know about hypothetical example (labelled)?
A correction changes a figure in an article. The database is updated, but the page is pre-rendered and the CDN caches HTML for an hour under a site-specific rule. Readers see the old figure for up to an hour after a rebuild, or indefinitely if no rebuild runs. Purging the URL without rebuilding just re-caches the…
What should you know about reader takeaway?
When an updated article still looks old, ask which copy you're looking at and what rule makes it fresh . Fix it at that layer (rebuild, purge or headers) and then verify the public URL.
References & sources
  1. Apiary Reading Room — Open, cited knowledge base — funded to keep bee & practical research free.
From the Apiary Reading Room. Opinion & editorial — not financial advice. We don't overclaim.
More from the Reading Room