When to Use Cache-Control: no-store for Live IP Checks

ColdProxy Team7 min read

Layered paper scene showing a fresh response passing a closed cache tray

Live IP checks call for Cache-Control: no-store when reused response bytes could make a changed public exit look unchanged.

The directive tells compliant HTTP caches not to save the response for later reuse.

Key Takeaways

  • no-store is a storage rule. It tells compliant HTTP caches not to retain the request or response for later reuse.
  • ColdProxy returned it in one dated check. On August 11, 2026, the public IPv4 My IP endpoint carried private, no-store, max-age=0.
  • An ETag leaves the storage rule unchanged. A validator does not grant a cache permission to save a response.
  • Sticky sessions can repeat an exit IP after a fresh response. Response reuse and proxy exit assignment require their own checks.

What Does Cache-Control: no-store Mean for a Live IP Check?

For a live IP check, no-store stops a compliant HTTP cache from saving the response and serving those saved bytes to a later request.

RFC 9111, HTTP Caching was published in 2022 and states the response rule directly: "a cache MUST NOT store any part of either the immediate request or the response". The same section applies the directive to private and shared caches.

An IP-check response records the public identity observed for one request. Reused bytes can leave the old exit identity on screen after a successful proxy-route switch. The resulting diagnosis may be wrong.

Consider an authorized QA session that switches proxy routes and opens the same IP-check URL again. An unchanged result leaves three possibilities: a cache reused an earlier response, no new request reached the endpoint, or a new request legitimately kept the same exit. The displayed address alone cannot identify which event occurred.

What Did ColdProxy's My IP Endpoint Return?

On August 11, 2026, we made three requests on public ColdProxy surfaces: one page GET to ColdProxy's My IP page, one direct GET to its IPv4 endpoint, and one HEAD request to that endpoint. The two endpoint requests used the same Fetch cache mode as the page code. IP, hostname, ASN, organization, and location values were discarded.

  • Tool page GET: HTTP 200 with the H1 What Is My IP Address?.
  • Endpoint GET: HTTP 200 with structured JSON and Cache-Control: private, no-store, max-age=0.
  • Other saved metadata: an ETag was present, while an Age header was absent.
  • Endpoint HEAD: HTTP 200 with zero body bytes. This HEAD check did not return the JSON exit-address value used by the My IP tool.

The sample did not query the IPv6 endpoint, execute the page in a rendered browser, send a conditional request, test a proxy route, or measure performance. It covers one public IPv4 endpoint under the tool's Fetch mode, so the result should not be extended to every ColdProxy response.

How Do private, no-store, max-age=0, and ETag Differ?

The observed response includes four cache controls or metadata fields:

  • private — shared-cache scope. A shared cache must not store the response. A private browser cache may store it when no stricter directive forbids storage.
  • no-store — storage prohibition. A compliant HTTP cache must not retain the request or response for later reuse.
  • max-age=0 — immediate staleness. If a response is stored, its freshness lifetime ends immediately, so ordinary reuse requires validation.
  • ETag — representation validator. A client can send its value with If-None-Match when validation is appropriate. Its presence does not grant storage permission.

MDN's HTTP caching guide draws the useful line: no-cache permits storage but requires validation before reuse, while no-store prevents storage. In the ColdProxy combination observed here, no-store makes the storage decision. The other fields retain their narrower roles.

Why Can an ETag Appear Beside no-store?

An ETag can appear beside no-store because response generation can emit validation metadata even when the cache policy forbids storage.

A generic response pipeline can produce both headers. Cache-Control remains the authority for storage permission, so the ETag gives no evidence that a cache saved the body. It would become useful only where a stored representation and a later conditional request make validation possible.

Our record contains the ETag-presence flag only. It excludes the validator value and any conditional-request result such as 304 Not Modified. This boundary protects the live value and avoids treating one response header as proof of behavior that was not tested.

Does no-store Clear an Old Cache Entry?

No. Existing cache entries remain under their prior lifecycle until eviction or explicit cache management removes them.

The MDN HTTP caching reference notes that no-store does not delete an already-stored response. A versioned URL uses a new cache key and leaves the old entry stored until its lifecycle ends.

Set the intended cache policy when the live identity endpoint first ships. A later no-store directive applies to new responses.

Does no-store Also Control the Browser Back/Forward Cache?

Browser back/forward cache, or bfcache, applies its own history-restoration eligibility and eviction rules alongside HTTP cache policy.

The HTTP Archive 2025 Performance chapter source reports that its 2025 analysis found Cache-Control: no-store on 23% of sites, up from 21% in 2024. The published shares differ by two percentage points.

In a 2023 public Chrome bfcache proposal, Fergal Daly and Kenji Baheux report Chrome measurements in which no-store prevented about 17% of mobile history navigations and 7% of desktop history navigations from using bfcache. Those percentages use history navigations as their denominator; they do not measure sites or live-IP endpoint traffic.

Treat HTTP storage and page-history restoration as two checks. Apply no-store to the storage decision, then test bfcache behavior independently when a restored document matters to the user experience.

Which Cache Directive Fits an IP-Check Endpoint?

Choose the directive by asking whether saved response bytes may answer a later request:

  • Use no-store when reuse would mislead the caller. A live identity response belongs here when an older body could hide a changed public exit.
  • Choose no-cache when storage is acceptable but reuse needs validation. The cache can retain a copy, then check with the origin before serving it again.
  • A positive max-age works when reuse for a known interval is acceptable. This fits content whose value remains useful throughout that interval.
  • Mark a response private when it belongs to one client. Add an explicit storage rule when the private cache must not retain it.

HTTP cache directives govern response storage and reuse. Check the service's policies for server logging and other retention, and inspect the proxy route for exit assignment. Do not present a no-store header as a general privacy promise.

Why Can a Proxy Result Still Look Stale?

When an address repeats, run the three-check stale-result test:

  • Response freshness: did an HTTP cache reuse old bytes?
  • Request execution: did a new request reach the identity endpoint?
  • Exit continuity: did the proxy legitimately keep the same exit for the session?

The gateway-versus-exit guide explains why the proxy hostname you connect to and the public address a destination sees are distinct values. The guide to rotating and sticky proxy sessions covers why a valid new request can keep one exit.

For a broader production check, use the proxy testing checklist and the connection verification guide. They cover reachability and observed identity without treating one repeated value as proof of a cache fault.

How Should You Verify a Live IP Check?

If you control a live identity endpoint, send Cache-Control: no-store when later requests require newly generated responses. Then verify request execution and compare the returned exit with the session and rotation behavior you expected.

ColdProxy's public IPv4 My IP endpoint returned private, no-store, max-age=0 in this dated observation. For an authorized testing workflow, compare ColdProxy plans and use the My IP tool to check the observed exit against the expected session and rotation behavior.

Frequently Asked Questions

How does no-store differ from no-cache?

no-store forbids a compliant HTTP cache from retaining the response, while no-cache allows storage but requires validation before reuse. That distinction matters when an endpoint must return a newly generated result: choose the directive based on whether storage itself is acceptable, not on how fresh the last response appears.

Does no-store guarantee a different exit IP on every request?

No. Proxy session and rotation settings determine exit assignment, so a new request may legitimately keep the same public address. Check the expected sticky or rotating behavior before diagnosing a cache problem, and confirm that the request actually reached the identity endpoint.

Does an ETag override no-store?

No. Under Cache-Control: no-store, an ETag remains validation metadata. The cache follows the storage permission in Cache-Control; the validator adds nothing to that permission. A generic response pipeline may emit both headers on the same response, so their coexistence does not show that a cache retained the body.

Does no-store prove that an endpoint keeps no logs?

No. HTTP cache directives govern storage and reuse by HTTP caches. Logging, analytics, security records, and other server-side retention follow independent policies. Read the service's published privacy terms for those details instead of treating a response header as a statement about every data-handling system.

Can a HEAD request verify a proxy exit address?

Not in our recorded check. The HEAD response had zero body bytes and did not provide the JSON exit-address value used by the My IP tool. Use GET when you need that value, then assess proxy reachability and session behavior. A HEAD response can still expose other metadata.

ColdProxy Team

ColdProxy Team

Content Team

The ColdProxy Content Team consists of proxy-service experts, developers, and technical writers dedicated to providing clear, accurate insights on web scraping, online privacy, and advanced proxy technologies.