Seeing "all sentinels specified in configuration are unreachable" error

Description

I am seeing the following error returned from my redis client: "all sentinels specified in configuration are unreachable" It is happening relatively infrequently, but often enough to cause issues with the application code. I do not see any corresponding errors in the redis server logs.

Expected Behavior

In theory I should never see this error.

Current Behavior

I get "all sentinels specified in configuration are unreachable" periodically (maybe 1% of the time). Do I need to implement retry logic?

Steps to Reproduce

I do not know how to reproduce

Context (Environment)

3 redis nodes in my redis cluster (master + 2 sentinels)

This error has caused the lookup of cached queries to fail, and redis Counters not to be reliable. Redis counters are used to bypass certain functionality in the app if a count reaches a certain threshold.

Version

github.com/redis/go-redis/v9 v9.3.0

Attempts to remediate

set redis client configuration to include:

    client = redis.NewClient(&redis.Options{
        Addr:           config.Local.Redis.Addr,
        MaxRetries:     5,
        PoolSize:       200,
        MaxActiveConns: 0,
    })

disable any other redis clusters (they all run on the same port)
Test in 3 completely different environments

redis / go-redis