## Summary As mentioned in the linked incident issue, the problem was a single `fe` node where we saw a large amount of network retries that contributed to apdex drops across all of the front-end services. I've opened up a support ticket with google https://console.cloud.google.com/support/cases/detail/29546343 , in the `gitlab-production` project to see if they can identify a host problem on their end. <!-- Give context about the investigation --> ## Related Incident(s) <!-- Note the originating incident(s) and link known related incidents/other issues. --> Originating issue(s): gitlab-com/gl-infra/production#6571