Dedicated proxy HTTPS check: 5–7 October 2026
We ran a small, fixed HTTPS check using three of our own Dedicated proxies reserved for testing. Each proxy was checked against five public registry websites on each of three UTC dates: 5, 6 and 7 October 2026.
We attempted 45 requests. Five requests from one proxy on 5 October were excluded because the execution environment did not have the required network access. A direct control request also failed in that restricted environment and returned HTTP 200 when network access was enabled. We kept those original observations and did not repeat or replace them.
The remaining 40 requests each returned HTTP 200. This describes the responses at the sampled times; it is not a measurement of continuous uptime or a guarantee for other destinations.
| UTC date | Requests attempted | Qualified observations | Environment exclusions | HTTP 200 / qualified observations |
|---|---|---|---|---|
| 5 October 2026 | 15 | 10 | 5 | 10/10 |
| 6 October 2026 | 15 | 15 | 0 | 15/15 |
| 7 October 2026 | 15 | 15 | 0 | 15/15 |
| Total | 45 | 40 | 5 | 40/40 |
How the check worked
The requests ran from a local Windows computer using cURL, through an authenticated HTTP proxy to HTTPS destinations. We used HEAD with HTTP/1.1, verified TLS, a six-second connection and total timeout, no redirects and no retries. Requests ran one at a time, with at least ten seconds between proxy batches.
The fixed destinations were RIPE, whois.kr, Registro.br, AFRINIC and APNIC. We requested response headers rather than page bodies. An HTTP 200 response does not establish that a browser session, login or application will work.
Across the 40 qualified observations, the median total cURL time was 1157.5 ms, with a range of 237–3533 ms. This includes the connection to the proxy, CONNECT, TLS negotiations and the HEAD response. It is not a measurement of proxy-only latency or download throughput.
Timing and limits
The first day's checks ran at approximately 17:22–17:25 UTC. The second and third days ran at approximately 20:36–20:39 UTC after the scheduled triggers arrived later than planned. The checks therefore did not all share the same time-of-day window. The underlying cause of the delayed trigger delivery was not independently established.
This is a small sample of reserved test proxies. We did not test continuous availability, download speed, SOCKS5, destination login, application usability or geographic location. We did not include a Semi-Dedicated comparison group, and the results should not be used to rank proxy products.
Private proxy addresses and authentication details are excluded from the study evidence. We retain anonymous observations, timestamps and the execution settings so that the reported counts and exclusions can be checked.
Anonymous observations
Download the anonymous CSV (45 observations, including five environment exclusions).
The excluded field is True for the five observations excluded from performance statistics; their exclusion_reason is sandbox-network-restriction. T01, T02 and T03 are anonymous proxy aliases. Destination IDs in the CSV map to these fixed public URLs: