Hello n8n team
The idea is:
Add support for Sentinel for the Redis used in queue mode and multi-main instances.
My use case:
We are hosting n8n version 1.123 on-prem on Kubernetes.
We currently use a simple redis setup with one master instance and replicas.
I think it would be beneficial to add this because:
In production HA environments, Sentinel is important for automatic master discovery and failover handling. Without Sentinel, Redis becomes a large single point of failure.
Any resources to support this?
Are you willing to work on this?
Definitely, currently this is one of the biggest pain points we have with deploying n8n onprem.
Target Setup:
Our target setup is to be able to configure Redis connection via Sentinel endpoints instead of a single host. We want to be able to use N Sentinel instances (behind a singular kubernetes service) and we should also be able to configure Auth and TLS if it is needed.
Both workers and main instances should be able to make use of Sentinel to discover the redis master.