In my last infrastructure update, I documented our challenges with storage as GitLab scales. We built a CephFS cluster to tackle both the capacity and performance issues of NFS and decided to replace PostgreSQL standard Vacuum with the pg_repack extension.
Full article