info: To determine the technical writer assigned to the Stage/Group associated with this page, see https://about.gitlab.com/handbook/product/ux/technical-writing/#assignments
| Node | An individual server that runs GitLab either with a specific role or as a whole (for example a Rails application node). In a cloud context this can be a specific machine type. | GitLab | instance, server |
| Site | One or a collection of nodes running a single GitLab application. A site can be single-node or multi-node. | GitLab | deployment, installation instance |
| Primary site | A GitLab site whose data is being replicated by at least one secondary site. There can only be a single primary site. | Geo-specific | Geo deployment, Primary node |
| Secondary site | A GitLab site that is configured to replicate the data of a primary site. There can be one or more secondary sites. | Geo-specific | Geo deployment, Secondary node |
| Geo deployment | A collection of two or more GitLab sites with exactly one primary site being replicated by one or more secondary sites. | Geo-specific | |
| Reference architecture | A [specified configuration of GitLab for a number of users](../reference_architectures/index.md), possibly including multiple nodes and multiple sites. | GitLab | |
| Failover | The entire process that shifts users from a primary Site to a secondary site. This includes promoting a secondary, but contains other parts as well. For example, scheduling maintenance. | Geo-specific | |