docs: add scaling section to deployment docs

Signed-off-by: Patrik Oldsberg <poldsberg@gmail.com>
2022-07-15 14:09:34 +02:00
parent a3b9b7f012
commit b745fa726c
3 changed files with 26 additions and 0 deletions
@@ -0,0 +1,24 @@
+---
+id: scaling
+title: Scaling Backstage Deployments
+sidebar_label: Scaling
+description: Scaling Backstage production deployments
+---
+
+There are several different methods for scaling Backstage deployments. The most
+straight-forward one is to simply deploy multiple identical instances and distributing
+incoming requests across them. The one requirement for this to work is that all instances
+need to share the same external resources, such as the database, and optional caching or
+search services. The Backstage backend plugins will coordinate through the database
+to share state and coordinate work.
+
+Another method for scaling Backstage deployments is to break apart the backend
+into multiple different services, each running a different set of plugins. This
+is a more advanced approach and requires you to be able to route requests to
+the appropriate backends based on the plugin ID. Both for ingress, but also
+internal traffic between Backstage backends, which is done by creating a custom
+implementation of the [PluginEndpointDiscover](../reference/backend-common.pluginendpointdiscovery.md) interface.
+
+Lastly, you can also replicate the Backstage deployments across multiple regions.
+This is not a pattern that there is built-in support for and typically only makes
+sense to do for individual backend plugins.