### Key Discussion Points

*   **Meeting Schedule:** The meeting on 2025-11-27 will be canceled due to a holiday in the US.
*   **On-Call Summary:** The past week included four pages for a Koji outage, mailing list ownership, a stuck Bodhi update, and a stuck Koji build. The signing queue also required two restarts.
*   **TCP Timeout Bug:** A major, long-standing TCP timeout bug appears to be resolved following a firewall cluster upgrade. This is evidenced by a dramatic drop in active connections from over 1 million to ~150k. A new, less severe issue related to Koji logs still exists.
*   **Outage Process:** There was a discussion about the recent increase in outages. A proposal was made to create a formal process for logging outages, publishing Root Cause Analysis (RCA) documents, and ensuring the status page is updated promptly during incidents.
*   **Monitoring (Nagios):**
    *   The `*.apps.ocp.fedoraproject.org` certificate expires in 22 days.
    *   There is a disk space issue on `vmhost-x86-02`.
*   **Monitoring (Zabbix):**
    *   High load on `pkgs01` is attributed to scraper activity. The average load over the last 7 days was 12.
    *   A discussion has been started on Discourse to address Zabbix alert noise.
*   **Forge Migration:**
    *   The team discussed the migration of infrastructure repositories (ansible, tickets) to the new GitLab-based forge.
    *   A significant concern is the lack of private ticket functionality, which is required for security issues. The interim solution will be to direct private reports to email.
    *   The migration is tentatively planned for the second week of December to avoid conflicting with the rdu-cc data center move.

### Action Items

*   **@seddik (saibug):** To chair the next meeting on 2025-11-20.
*   **@nirik:** To be on-call for the upcoming week (starting 2025-11-14).
*   **@james:** To renew the `*.apps.ocp.fedoraproject.org` SSL certificate.
*   **@gwmngilfen:** To adjust the Zabbix load alert threshold for `pkgs01` to trigger around a load of 6.
*   **@gwmngilfen:** To modify the postfix queue monitoring in Zabbix to alert on messages that are stuck over time, rather than on a simple queue count.
*   **@nirik:** To move `kojipkgs` back to using port 80 (Varnish).
*   **@nirik:** To close the old TCP timeout ticket and open a new one specifically for the remaining Koji 502 error.
*   **@nirik:** To start a discussion thread to propose a formal process for outage logging and RCAs.
*   **@nirik:** To update the forge migration ticket with a concrete plan and timeline.