ops: Revisit how to SSH into Gitaly hosts

During the recent incident (INC-649), responders encountered difficulties accessing the affected Gitaly host after it was migrated to the new project structure. This significantly delayed our resolution time as incident responders spent valuable time figuring out how to SSH into these nodes to trigger the startup script necessary for rollback.

Eventually, our SREs gained the access from SSH configs magically persisted somewhere in their machines. But this this documentation gap presents a critical operational risk. When responding to Gitaly incidents, direct SSH access is often necessary to diagnose problems, collect performance data, or execute recovery procedures.

We need to create clear documentation that outlines how to SSH into Gitaly hosts in the new project structure, highlighting any differences from the previous access methods. This documentation should be easily accessible to all SREs and relevant engineering teams to ensure quick response during future incidents.