Replica

As explained in Kafka for Architects

A replica is a copy of a Kafka partition's records maintained on a broker for redundancy and availability.

A replica is a copy of a Kafka partition maintained on a broker. Each partition has one leader and one or more followers, allowing copies to be distributed across machines for redundancy and availability. section 2.3.1

Replication introduces both resilience and operational trade-offs:

• Producers write to the partition leader; followers asynchronously pull new records, so replicas can temporarily lag. section 2.3.1
• An in-sync replica (ISR) has caught up with the leader sufficiently to participate in reliable writes. section 2.3.1
• The replication factor sets the number of copies, while min.insync.replicas sets how many caught-up replicas must be available for a successful write. section 2.3.2
• A single replica does not provide high availability; a commonly recommended resilience profile uses replication factor 3 and min.insync.replicas 2. section 2.3.2

Figure 2.10 Each partition has one replica designated as the leader, while the others serve as followers.

Use the leader–follower diagram to compare replica roles and their placement across brokers. section 2.3.1

How Kafka for Architects covers it

How deep it goes

Defines placement, asynchronous lag, ISR status, and write thresholds, while deferring detailed client reliability strategies.

Examples

Uses leader–follower placement, asynchronous follower catch-up, and a resilience profile of replication factor 3 with min.insync.replicas 2.

What it assumes

Assumes Kafka partitions and brokers, while defining leader, follower, and in-sync replica relationships.

What you can do after

Configure replication factor and min.insync.replicas to balance availability, durability, and write acceptance.

How it connects

Connects replicas to Kafka partitions, brokers, leaders, followers, in-sync replicas, replication factor, and min.insync.replicas.

liveBookmanning.com

Read Kafka for Architects on liveBook Get the book at manning.com

Also explained in

• 

  liveBookmanning.com
  Elasticsearch in Action, Second EditionA duplicate copy of a primary shard maintained on a different node for redundancy and read availability.
• 

  liveBookmanning.com
  Platform Engineering on KubernetesA replica is an additional running instance of a service or application component that provides redundancy and capacity.
• 

  liveBookmanning.com
  Apache Kafka in ActionA Kafka replica is a copy of a partition maintained on a broker for fault tolerance and high availability.
• 

  liveBookmanning.com
  Kubernetes in Action, Second EditionAn application instance is a running replica of an application managed by Kubernetes.
• 

  liveBookmanning.com
  ScyllaDB in ActionA replica is a copy of database data maintained on another node for availability, resilience, or read serving.
• 

  liveBookmanning.com
  Learn Docker in a Month of Lunches, Second EditionA replica is an independently running instance of a workload maintained to provide capacity or redundancy.
• 

  liveBookmanning.com
  Acing the Certified Kubernetes Administrator Exam, Second EditionA replica is an independently running copy of a pod or application instance that contributes to workload capacity and redundancy.
• 

  liveBookmanning.com
  Bootstrapping Microservices, Second Edition: With Docker, Kubernetes, GitHub Actions, and TerraformA replica is an additional running instance that duplicates the functionality of another service or component.
• 

  liveBookmanning.com
  Server-Side WebAssemblyA replica is an independently running copy of a component or service deployed to handle workload alongside other copies.
• 

  liveBookmanning.com
  LLMs in ProductionA replica is an additional running instance of a service or workload that provides capacity or redundancy.
• 

  liveBookmanning.com
  LatencyA replica is a server or other system member that maintains a copy of data or state replicated from a primary.

Compare how 12 books explain replica →

Related concepts

• horizontal scalingused by · Horizontal scaling is adjusting system capacity by adding or removing computing nodes or equivalent instances that share or distribute the workload.
• replicationproduced by · Replication is the creation or maintenance of multiple copies of data, state, or functionality across separate locations or components.
• uptimeproduces · Uptime is the duration or proportion of a specified period during which a system or service is operational and available to provide its intended function.
• primary nodecontrasts with · A primary node is a designated authoritative or coordinating member of a distributed system, typically responsible for accepting operations or coordinating replicated state.
• load balancerused by · A load balancer is a hardware or software component that distributes incoming network traffic or requests among multiple destinations, such as servers or service instances.
• in-sync replicaincludes · An in-sync replica is a Kafka partition replica sufficiently caught up with its leader to remain eligible for failover.

This explainer was generated by Manning's knowledge graph from the text of Kafka for Architects. The section links point to the parts of the book it draws on.