Skip to content
Runbooks

Cloudflare is Down

Since Cloudflare is our entry point, when it’s down, most of our services will be unavailable to users. Our primary focus during these incidents is communication, monitoring, and preparing for recovery.

Possible Actions

Section titled “Possible Actions”
  • Communicate with Cloudflare:
    • Post in shared Slack channel: #gitlab-cloudflare-support
    • Add the :ticket: emoji to open a support case
    • Request ETA and ongoing updates
  • Engage CMOC to update the GitLab status page. Consider the following template:
    • GitLab.com is down due to an upstream provider outage
    • GitLab.com is inaccessible at the moment due to an upstream provider outage. We are currently investigating potential mitigation options.
  • If only a particular feature or rule is affected by the downtime, consider disabling it temporarily to allow access to the site in the meantime.
    • Always create a follow-up issue to re-enable whatever was disabled.
    • A terraform apply in the gprd environment will reset the rules, which can be used as a post-incident fix.