How to minimize service unavailability during Raft leader failure? Industry best practices for reducing failover time
## Context I'm working on a distributed consensus system based on the Raft protocol. Currently, when the leader node fails, our service experiences unavailability during the period of:
1 Follower timeout detection (waiting for heartbeat timeout)
2 (candidate election and vote
collection) This unavailability window typically ranges from 2-5
seconds depending on configuration, which impacts our SLA
requirements.
## Question **What are the industry-proven techniques to minimize this unavailability window?**