diff --git a/docs/upgrading.rst b/docs/upgrading.rst
index a9c50bc90ba23966b790a09e43da91eda9f85597..e53f51b4f29d1970ff9f55e74f3bf178204c4cbc 100644
--- a/docs/upgrading.rst
+++ b/docs/upgrading.rst
@@ -1,6 +1,35 @@
 Upgrading
 =========
 
+Upgrading to 0.8.0 (not released yet)
+-------------------------------------
+
+When upgrading to version 0.8.0 you'll get the newest versions of several
+database helm charts that have changed significantly. Because they have, you'll
+have to manually remove the following statefulsets (note: as long as you haven't
+overwritten any of the `persistence:` variables for those charts, you can do
+this without risk. If you have, make sure to back up your PVCs).
+
+```
+kubectl delete statefulset -n oas-apps nc-mariadb nc-redis-master nc-rabbitmq nc-postgresql
+```
+
+If your Nextcloud helmrelease gets stuck in a non-ready state, you can run this
+to force re-reconciliation:
+
+```
+flux suspend helmrelease nextcloud -n oas-apps
+flux resume helmrelease nextcloud -n oas-apps
+```
+
+If aftewards your ``nc-nextcloud-xxxxxxxxxx-xxxxx`` pod is stuck in the
+``Initializing`` phase, you can delete it to make sure its replacement connects
+to the new databases:
+
+```
+kubectl delete pod -n oas-apps nc-nextcloud-xxxxxxxxxx-xxxxx
+```
+
 Upgrading to 0.7.0
 ------------------