InditexTech · albertompe · Nov 24, 2025 · Nov 24, 2025 · Nov 24, 2025
diff --git a/README.md b/README.md
@@ -1,4 +1,5 @@
-# ![RedKey Operator icon](docs/images/badge.png) RedKey Operator for Kubernetes
+![RedKey Operator icon](docs/images/redkey-logo-50.png)
+# RedKey Operator for Kubernetes
 
 A **RedKey Cluster** is a key/value cluster using either [Redis Official Image](https://hub.docker.com/_/redis) or [Valkey Official Image](https://hub.docker.com/r/valkey/valkey/) images to create its nodes (note that all cluster nodes must use the same image).
 
@@ -12,13 +13,13 @@ RedKey operator is built using [kubebuilder](https://github.com/kubernetes-sigs/
 
 - RedKey Cluster creation
 - Cluster scaling up and down
-- CLuster upgrading
+- Cluster upgrading
   - Update node pods image
   - Update Redis configuration
   - Update node pods resources
 - Ensure cluster health
 - Slots allocation
-- Persistence
+- Ephemeral cluster (pure cache-like behavior) or using persistence
 - RedisGraph support
 
 ## Documentation
@@ -27,7 +28,11 @@ Refer to [operator guide](./docs/operator-guide/toc.md) to have an overview of t
 
 If you are a developer, you'll find interesting information in the [developer guide](./docs/developer-guide.md).
 
-Learn about [RedKey cluster status](./docs/redkey-cluster-status.md).
+Learn about [RedKey Cluster Status and Substatus](./docs/redkey-cluster-status.md).
+
+Discover [RedKey Robin](./docs/redkey-robin.md).
+
+The importance of the [purgeKeysOnRebalance](./docs/purge-keys-on-rebalance.md) parameter.
 
 ## Versions
 

diff --git a/docs/developer-guide.md b/docs/developer-guide.md
@@ -349,7 +349,7 @@ Before redeploying the webhook code using `make debug-webhook` you can delete th
 Delete the operator and all associated resources with:
 
 ```shell
-make delete-rdcl
+make delete-rkcl
 make undeploy
 make uninstall
 kubectl delete ns redkey-operator
@@ -360,10 +360,14 @@ kubectl delete ns redkey-operator
 You can deploy a sample RedKey Cluster from `config/samples` folder running:
 
 ```shell
-make apply-rdcl
+make apply-rkcl
 ```
 
-This will apply the manifest file to create an single node ephemeral RedKeyCluster object with `purgeKeysOnRebalance` set to **true**.
+This will apply the manifest file to create a three nodes ephemeral RedKeyCluster object with `purgeKeysOnRebalance` set to **true**.
+
+As [Redkey Robin](https://github.com/InditexTech/redkeyrobin) image is required to being able to deploy the RedKey cluster, you should ensure the image is available in your K8s cluster. You'll find how to build the image from the GitHub repository.
+
+In order to set the image name to be used when creating your local RedKey cluster you can sen the environment variable `IMG_ROBIN`. The value set to this variable will be used to kustomize the manifests that will be applied to create the sample RedKeyCluster.
 
 ## Tests
 

diff --git a/docs/images/redis-status-creation-meet-rebalance.png b/docs/images/redis-status-creation-meet-rebalance.png
diff --git a/docs/images/redis-status-error.png b/docs/images/redis-status-error.png
diff --git a/docs/images/redis-status-scale.png b/docs/images/redis-status-scale.png
diff --git a/docs/images/redis-status-upgrade.png b/docs/images/redis-status-upgrade.png
diff --git a/docs/images/redkey-cluster-initialization.png b/docs/images/redkey-cluster-initialization.png
diff --git a/docs/images/redkey-cluster-substatus-scalingdown-fast.png b/docs/images/redkey-cluster-substatus-scalingdown-fast.png
diff --git a/docs/images/redkey-cluster-substatus-scalingdown-slow.png b/docs/images/redkey-cluster-substatus-scalingdown-slow.png
diff --git a/docs/images/redkey-cluster-substatus-scalingup-fast.png b/docs/images/redkey-cluster-substatus-scalingup-fast.png
diff --git a/docs/images/redkey-cluster-substatus-scalingup-slow.png b/docs/images/redkey-cluster-substatus-scalingup-slow.png
diff --git a/docs/images/redkey-cluster-substatus-upgrading-fast.png b/docs/images/redkey-cluster-substatus-upgrading-fast.png
diff --git a/docs/images/redkey-cluster-substatus-upgrading-slow.png b/docs/images/redkey-cluster-substatus-upgrading-slow.png
diff --git a/docs/images/redkey-logo-50.png b/docs/images/redkey-logo-50.png
diff --git a/docs/images/redkey-logo.png b/docs/images/redkey-logo.png
diff --git a/docs/images/redkey-operator-robin.png b/docs/images/redkey-operator-robin.png
diff --git a/docs/purge-keys-on-rebalance.md b/docs/purge-keys-on-rebalance.md
@@ -0,0 +1,31 @@
+# Key purge (purgeKeysOnRebalance)
+
+When using RedKey Cluster as a cache (`ephemeral` parameter set to **true**), we have the option to tell the Operator whether we want to keep the keys when performing scaling or upgrade operations, or whether we want to discard them.
+
+When using RedKey Cluster as a cache (ephemeral parameter set to true), we have the option to tell the Operator whether we want to keep the keys when performing scaling or upgrade operations, or whether we want to discard them.
+
+Scaling or upgrade operations have a **high computational and time cost**, as they involve resharding or moving slots (and their keys) between nodes of the cluster itself.
+
+In order to speed up these operations, a fast and slow mode has been implemented.
+
+More details about the Status and Sustatus for these operations can be found [here](.redkey-cluster-status.md).
+
+## Fast operations
+
+The fast mode of scaling and upgrade operations implies a **total loss of the keys** stored in the RedKey Cluster.
+
+In this mode, the Operator is actually recreating the RedKey Cluster, recreating the pods that make it up and relying on Robin to meet the new nodes, distribute the slots, and ensure that it remains in a healthy state.
+
+The cluster will be properly scaled or have the updated configuration.
+
+This is the fastest way to perform scaling and upgrade operations.
+
+Quick operations will only be applied if the RedKey Cluster is set to `purgeKeysOnRebalance` set to **true**.
+
+## Slow Operations
+
+If we want to preserve keys during RedKey Cluster upgrades and scales, then we need to use slow operations. To do this, we must configure the cluster with `purgeKeysOnRebalance` set to **false**.
+
+In this case, the Operator makes sure to move the slots (and their keys) from one node to another node or nodes within the cluster itself, before recreating that node or discarding it.
+
+The necessary reharding operations involve a high cost, which is impacted if the cluster receives a lot of activity from clients.