Added note about kafka not being master, and FAQ entry about schema b…

…ackup (#537)

* Added note about kafka not being master, and FAQ entry about schema backup

* Don't mention on-prem kafka at all

docs/persistence/kafka/README.md

@@ -1,13 +1,24 @@
 # Kafka
 
-NAV uses Aiven hosted Kafka, the on-premise version of Kafka has been deprecated since 2021. Start using Kafka by creating a `Topic` resource in one of our Kubernetes clusters.
+NAV uses Aiven hosted Kafka.
+Start using Kafka by creating a `Topic` resource in one of our Kubernetes clusters.
 
-Upon defining a Topic, [Kafkarator](https://github.com/nais/kafkarator) will create the Topic in one of the Kafka _pools_. A pool is a highly available, replicated Kafka cluster running at Aiven. After the topic is created, Kafkarator will add relevant users to the topic's access control list \(ACL\).
+Upon defining a Topic, [Kafkarator](https://github.com/nais/kafkarator) will create the Topic in one of the Kafka _pools_.
+A pool is a highly available, replicated Kafka cluster running at Aiven.
+After the topic is created, Kafkarator will add relevant users to the topic's access control list \(ACL\).
 
-When an application that uses Kafka is deployed, [Aivenator](https://github.com/nais/aivenator) will orchestrate generation of users and credentials. These credentials are made available to applications through a _Secret_ in the relevant team namespace. This secret is automatically mounted by Naiserator into application pods as environment variables and files.
+When an application that uses Kafka is deployed, [Aivenator](https://github.com/nais/aivenator) will orchestrate generation of users and credentials.
+These credentials are made available to applications through a _Secret_ in the relevant team namespace.
+This secret is automatically mounted by Naiserator into application pods as environment variables and files.
 
 To get started, see [accessing topics from an application](manage_topics.md#accessing-topics-from-an-application).
 
+!!! note "Backups and recovery"
+    Kafka as a system is highly durable, and is designed to be able to keep your data safe in the event of a failure.
+    This requires a properly configured replication factor for your topic, and that your clients use the appropriate strategy when sending messages and committing offsets.
+    Even so, our recommendation is that Kafka should not be the master of your data, and you should have the ability to restore your data from some other system.
+
+
 ## Status and roadmap
 
 * Release status: Generally Available

docs/persistence/kafka/faq.md

@@ -1,9 +1,6 @@
 
 # FAQ/Troubleshooting
 
-!!! warning
-    This feature applies only to _Aiven hosted Kafka_. On-premises Kafka is deprecated, and creating new topics on-premises was disabled summer 2021. For on-premises Kafka, see [on-premises Kafka documentation](https://confluence.adeo.no/display/AURA/Kafka).
-
 ## Why do I have to specify a pool name if there is only `nav-dev` and `nav-prod`?
 
 Custom pools might be added in the future, so this is done to avoid changing that part of the API.
@@ -19,3 +16,8 @@ Make sure you added the application to `.spec.acl` in your `topic.yaml`.
 ## I get the error _MountVolume.SetUp failed for volume "kafka-credentials" : secret ... not found_
 
 Check the status of the `AivenApplication` resource created by Naiserator to look for errors.
+
+## Are Schemas backed up?
+
+Aiven makes backups of configuration and schemas every 3 hours, but no topic data is backed up by default.
+See the [Aiven documentation](https://docs.aiven.io/docs/products/kafka/concepts/configuration-backup) for more details.