Greetings,
In regards to https://shinyproxy.io/documentation/shinyproxy-operator/kubernetes/#note-on-redis-and-valkey, I’d like to share our experience with switching from Redis to Valkey in our ShinyProxy environment on Kubernetes.
In short, we’ve been running ShinyProxy with Valkey in production for about a month now and have not experienced any issues.
Our environment:
- ShinyProxy version: 3.1.1
- ShinyProxy Operator version: 2.1.0
- Kubernetes version: 1.28
- Upgrade from Redis version: 7.2.4
- Upgrade to Valkey version: 8.1.3
- We use Helm charts for deploying all components to our Kubernetes cluster
- We use Redis/Valkey as a storage backend for session and app persistence across ShinyProxy server replicas
We did the migration as following:
- Deploy Valkey to the Kubernetes cluster and ensure everything is running fine
- Stop all running ShinyProxy applications
- Update ShinyProxy sentinel connection string in the ShinyProxy Kubernetes resource
- Wait for the ShinyProxy Operator to promote new ShinyProxy server replica set as leader
- Test and ensure that leader election on consecutive changes still works as expected
- Remove old Redis nodes from the cluster
We recently did an upgrade from Valkey 8.1.3 to 8.1.4 which also went without issues.