Leader Election

← Back to Distributed System Patterns

A pattern for choosing one node among many to serve as the coordinator/leader for a task. The leader handles writes, coordination, or scheduling while followers serve reads or act as standby. Implemented using consensus services (ZooKeeper, etcd) or built-in to consensus algorithms (Raft).

Key Properties

• Election Mechanism
• Leader Lease
• Failover

Related

• Raft (includes built-in leader election)
• Single-Leader Replication (uses leader election for failover)

---

distributed-systems patterns