Problem/Motivation

My daily scheduled actions on GHA have recently started failing frequently with the following error. Any idea what might be causing this?

PHP Fatal error: Uncaught GuzzleHttp\Exception\ConnectException: cURL error 35: OpenSSL SSL_connect: SSL_ERROR_SYSCALL in connection to www.drupal.org:443 (see https://curl.haxx.se/libcurl/c/libcurl-errors.html) for https://www.drupal.org/api-d7/node.json?type%5B0%5D=project_module&type%5B1%5D=project_theme&type%5B2%5D=project_core&field_project_has_releases=1&field_project_type=full&sort=field_project_machine_name&page=119 in /home/runner/work/ddqg/ddqg/vendor/guzzlehttp/guzzle/src/Handler/CurlFactory.php:275

https://github.com/mxr576/ddqg/actions/runs/16262130198/job/45909538532

Steps to reproduce

Proposed resolution

Remaining tasks

User interface changes

API changes

Data model changes

Comments

mxr576 created an issue. See original summary.

drumm’s picture

Status: Active » Postponed (maintainer needs more info)

We’re not aware of any general issues which would interrupt SSL connections.

We would need some more details to attempt to help, especially if it is intermittent.

mxr576’s picture

Status: Postponed (maintainer needs more info) » Active

Something is definitely ongoing, but I do not have more information other than what can be seen in the GitHub Actions failures. There have been many recently, most of them with the reported error, and others with:

 RuntimeException: Failed to fetch page 261. Reason: "cURL error 56: Recv failure: Connection reset by peer (see https://curl.haxx.se/libcurl/c/libcurl-errors.html) for https://www.drupal.org/api-d7/node.json?type%5B0%5D=project_module&type%5B1%5D=project_theme&type%5B2%5D=project_core&field_project_has_releases=1&field_project_type=full&sort=field_project_machine_name&page=261". in /home/runner/work/ddqg/ddqg/src/Infrastructure/DrupalOrg/DrupalOrgApi/DrupalOrgApiRepository.php:187
Stack trace:

https://github.com/mxr576/ddqg/actions/workflows/generate-no-release-ver...

This can also be seen in the action logs when the action runs, so I assume some infrastructure/WAF logs could be analyzed for anomaly.

drumm’s picture

Status: Active » Postponed (maintainer needs more info)

Both issues smell like they could be an issue anywhere from GitHub actions to our CDN, Fastly, and any of the internet in between; and there is also a portion of the internet between the CDN and our origin servers. We do not have detailed logs at all levels, and there’s not an efficient way to sift through the logs for connections from GitHub actions. We can engage Fastly support, but I also can’t expect them to be able to track down requests without details or reproducibility.