GitHub Pages requests intermittently hang without an HTTP response #209988
Replies: 7 comments 1 reply
|
Additional diagnostic evidence from Microsoft Edge NetLog I captured an Edge NetLog while reproducing the intermittent hang on https://www.cumnorhurst.org.uk/management/. The log shows: The browser established an IPv6 TCP connection to 2606:50c0:8002::153:443 and successfully negotiated TLS 1.3 and HTTP/2. The existing HTTP/2 connection successfully served several earlier requests, including /about/, /site.css, /volunteering/ and /events/. At approximately 13:22:33 BST, Edge sent an HTTP/2 ping and a GET request for /management/ over that existing connection. No incoming socket data is recorded during the following 10 seconds. At approximately 13:22:43 BST, the HTTP/2 session closed with ERR_HTTP2_PING_FAILED. Edge retried the request, which then received HTTP 200 from GitHub Pages in approximately 313 ms. The log therefore points to a previously working, reused HTTP/2 connection becoming unresponsive, followed by a successful retry on a replacement connection. It does not establish whether the cause is at the CDN/server, in the network path, or elsewhere. This is consistent with the intermittent hangs reported in Firefox and Edge, including by another user on a separate internet connection. Can anyone advise whether this pattern is known to occur with GitHub Pages or its CDN, and what further evidence would help identify the cause? |
|
The Edge NetLog narrows this substantially: the request succeeded after a replacement connection, so the page content and DNS are unlikely to be the trigger. For a clean comparison, test the same URL in a fresh browser process and with As a temporary client workaround, a new connection (private window, browser restart, or switching networks) should avoid the stale session; disabling HTTP/2 can be used only as a diagnostic because it changes browser-wide behavior. The strongest escalation packet is the HAR/NetLog plus exact UTC timestamps, URL, protocol, address family, and a successful retry from the same machine. If this diagnosis helps, please click Mark as answer. |
|
The Edge NetLog narrows this substantially: the request succeeded after a replacement connection, so the page content and DNS are unlikely to be the trigger. For a clean comparison, test the same URL in a fresh browser process and with As a temporary client workaround, a new connection (private window, browser restart, or switching networks) should avoid the stale session; disabling HTTP/2 can be used only as a diagnostic because it changes browser-wide behavior. The strongest escalation packet is the HAR/NetLog plus exact UTC timestamps, URL, protocol, address family, and a successful retry from the same machine. If this diagnosis helps, please click Mark as answer. |
|
The issue now appears more frequent. A second Edge NetLog capture again shows ERR_HTTP2_PING_FAILED after a 10-second timeout on a reused HTTP/2 session. The associated socket log records outgoing data when the ping and GET are sent, but no incoming socket data during the timeout, followed by socket closure. The request subsequently retries successfully. The same general problem occurs in Firefox and has also been reported by a user on a separate ISP. This also occurs on the default github.io hostname, not only the custom domain. |
|
The repeat capture makes the case stronger: this is a reused HTTP/2 session failure that reproduces across browsers, domains, and an ISP, with a successful retry on a new connection. At this point the highest-value next step is a GitHub Support report with the two NetLogs, exact UTC timestamps, affected hostnames, IPv4/IPv6 details, and the fact that the replacement connection returns 200. Keep one |
|
thanks for responding. during the period when firefox was stalled, i ran the following and successfully obtained the result i'm guessing that curl wasn't trying to use the same "old" connection and so didn't hit the same issue as the browser |
Uh oh!
There was an error while loading. Please reload this page.
🏷️ Discussion Type
Question
💬 Feature/Topic Area
Code Search and Navigation
Body
I’m experiencing intermittent failures when accessing my GitHub Pages website. GET requests sometimes hang without returning an HTTP response, leaving the previous page displayed in the browser.
Affected website
https://www.cumnorhurst.org.uk/https://microsonicsolutions.github.io/The site is hosted using GitHub Pages. I temporarily configured GitHub Pages to publish the CumnorHurst site directly at
microsonicsolutions.github.ioto investigate whether the custom domain was responsible. The site initially loaded successfully, but the same intermittent failure occurred again. This suggests the problem is not limited to the custom domain.Browser behaviour
I use Firefox on Windows. When I click a link to an affected page, the previous page remains displayed as though the link has not been followed. Firefox’s Network panel shows the corresponding GET request remaining incomplete, with no HTTP status or response data.
Eventually, the tab’s loading spinner stops, but the requested page has not loaded and the previous page remains visible. After the failure, the address bar still shows
microsonicsolutions.github.io, rather than the requested path such as/about/. I also noticed that a padlock appeared beside the hostname around the time the spinner stopped. I mention this as an observation, not as evidence of a specific TLS or HTTP failure.HAR evidence
I captured a HAR file from Firefox 157.0.1. One failed request was:
https://microsonicsolutions.github.io/about/0I understand that these HAR fields do not establish exactly where the request is failing, but they show that Firefox did not record a normal HTTP response for this navigation.
Tests already performed
These tests have not reproduced the browser failure consistently, so I suspect the issue may involve an intermittent interaction with the browser, connection handling, or the GitHub Pages/CDN delivery path. I cannot yet identify the cause.
Request for help
Could someone advise how best to investigate this, or whether there is a known issue affecting GitHub Pages requests that hang without returning an HTTP response?
In particular, I would appreciate advice on what additional diagnostics would help distinguish a browser/protocol issue from a GitHub Pages or CDN/edge issue.
I can provide the HAR capture privately if there is an appropriate way to do so. I have not attached it publicly because HAR files can contain sensitive request information.
Thank you.
All reactions