Why are 3 Availability Zones Better than 2? MongoDB Replica Sets Explained

Added: 2026-05-16

[00:00:00]Why does MongoDB practically insist on five nodes across three availability zones? It's not about overprovisioning.

[00:00:08]It's about the math of major. If you run a synchronous replication across only two sides, one primary and one secondary. A single network euphability to remain protected. or you decide to continue with unprotected rights and if the failure becomes a disaster then you lose data. In a distributed system you need a majority to elect a primary and acknowledge rights over zones. You can deploy five nodes, one primary and four secondaries. If the entire region goes dark, you still have three nodes, one primary and two secondaries, always at least two secondaries. The primary is selected by the majority within seconds and zero data loss. But what if you are stuck with only two data centers? The disaster recovery site may be left without a major edit. You can use a five node split with readonly nodes in the disaster recovery data center and this requires a manual post reconfiguring an outage with possible data loss. Two zones gives you a page of call. Three zones gives you resilience. The minimal configuration is three nodes. And the gold standard for zero drama distributed database is five nodes.