---
title: "Redis cluster fault tolerance"
sidebar_label: "Cluster fault tolerance"
sidebar_position: 12
description: "Role of the master node and replicas in a Redis cluster, node allocation in a cluster, how to change the number of replicas"
---
import ClusterFaultTolerance from '@site/i18n/en/docusaurus-plugin-content-docs/current/_partials/managed-databases/common/cluster-fault-tolerance.mdx'
import Formbricks from '@theme/MDXComponents/Formbricks'
# Redis cluster fault tolerance
By default, a Redis Managed Database cluster consists of a single [master node](#master-node). To provide cluster fault tolerance, you need to add [replicas](#replicas). In multi-node clusters, one node will be the master node, while the others will be replicas.
:::note
For example, if you have created a three-node cluster, the cluster will contain one master node and two replicas. If you add more nodes, they will be replicas.
:::
When adding replicas, consider the [node allocation type](#node-allocation-type) in the cluster.
## Master node \{#master-node}
By default, a cluster consists of one primary node, the master node. When connected to the master node, all read and write operations are available.
All data changes on the master node are replicated to the replicas. The replication process does not affect the performance of the replicas or the master node.
## Replicas \{#replicas}
Replicas are full copies of the master node. They are available for read-only data access.
Having replicas in a cluster ensures its fault tolerance — if the master node stops working, its role will be transferred to one of the replicas, and the cluster will continue to operate normally. When the master node is restored, it will take on the role of a replica. Note that the master node address will change.
For a cluster with replicas, an [SLA](https://files.selectel.ru/docs/ru/conditions-cloud-platform.pdf) applies — we guarantee 99.95% write availability and 99.99% read availability.
If a cluster has no replicas, it will be unavailable until the master node is restored. Data will not be lost during this process.
We recommend [creating fault-tolerant clusters](/managed-databases/redis/create-cluster.mdx) with replicas or [adding replicas](#change-number-of-replicas) to existing clusters.