Hoping someone can shed some light.
I am setting up a Redis Sentinel cluster as per configuration below:
- Node 1 - Master Redis + Sentinel 1
- Node 2 - Slave Redis + Sentinel 2
- Node 3 - Sentinel 3 (to maintain quorum)
The way we will use Redis is to setup a Redis instance per application ie. I will have a Redis instance/service listening on port 6380 for App 1, another Redis instance/service listening on port 6381 for App 2 etc on the same Master and Slave.
My question is, for the Redis Sentinel configuration, do we need a Redis Sentinel configuration setup to listen on 26380 for App 1 ,and then another configuration to listen on 26381 for App 2, etc.?
Thanks in advance!
A Sentinel deployment (made up of 3+ Sentinel instances) can manage multiple Redis deployments (made up of 1 master and 0 or more replicas). The only matter to consider is scalability - the more Redis deployments a single Sentinel deployment manages, the more CPU/RAM/NW resources it will need and eventually you'll need to scale out. That said, based on internet gossip, I heard that a single Sentinel deployment can handle 100s of Redis instances without major impact. YMMV, so please test before rushing into 10K deployments in production though :)