We have setup redis HA using sentinel on a kubernetes 3 node cluster using the bitnami image. Pls note the following configs were changed from 60000/180000 to the following
REDIS_SENTINEL_DOWN_AFTER_MILLISECONDS set to 6000
REDIS_SENTINEL_FAILOVER_TIMEOUT set to 18000
When there is a failure of the master node. Kubernetes brings up the instance quite quickly. However the sentinel does not failover to the new master quick enough. Consequently the write operations fail for 5-7 minutes. Can someone tell what can be done to switch over to new master immediately on a failover. Also where can we get more documentation on the above flags and how they impact our solution ?