Comparison

Updated October 25, 2026

· ~2 min read

Amazon MSK vs Strimzi on EKS vs Confluent: Running Kafka on AWS in 2026

Amazon MSK
Kafka
Strimzi
Confluent
KRaft
Streaming

Kafka 4.0 removed ZooKeeper entirely in favour of KRaft, and the managed options caught up through 2025 and 2026: Amazon MSK added Kafka 4.0 in May 2025, KRaft for Express brokers from Kafka 3.9, and in August 2026 in-place migration of existing clusters from ZooKeeper to KRaft. Express brokers, launched at the end of 2024, offer up to three times the throughput per broker, scale up to 20 times faster and recover 90% faster than standard brokers, according to AWS.

On Kubernetes, the Strimzi operator — a CNCF project — runs Kafka the same way on any cluster. And Confluent Cloud remains the full managed platform with its own ecosystem.

At a glance

MSK Express brokers

MSK standard brokers

Strimzi on EKS

Confluent Cloud

Storage

Managed, no sizing

EBS you size

EBS you size and manage

Managed

Scaling

Up to 20x faster than standard

Partition reassignment

Partition reassignment you run

Managed

Pricing

Broker hours plus per-GB written

Broker hours plus storage

Nodes, volumes and your time

Usage-based, vendor pricing

KRaft

From Kafka 3.9

Supported; in-place ZooKeeper migration since Aug 2026

Yes

Yes

Portability

AWS

AWS

Any Kubernetes

Multi-cloud

Operations

Lowest

Low

High

Lowest

Do you need Kafka at all?

Kafka is the right tool for high-throughput event streams that several consumers read independently and replay. It is the wrong tool for a work queue between two services, a handful of events a second, or fan-out notifications — SQS, SNS and EventBridge do those for a fraction of the cost and with nothing to operate. A surprising number of Kafka clusters on AWS carry less traffic than a single SQS queue handles comfortably.

Which should you use?

Question 1 of 2

Do several independent consumers need to read and replay a high-volume stream?

Our verdict

On AWS, use MSK — Express brokers for most new clusters, since they remove storage sizing and scale far faster — unless you need Confluent's ecosystem (managed connectors, stream processing, governance) badly enough to pay for it, in which case Confluent Cloud. Run Strimzi on EKS when Kafka must run the same way outside AWS, or when you already operate Kafka well and want full control of versions and configuration. If you are still on ZooKeeper-based MSK, plan the in-place KRaft migration now; Kafka 4.x does not have ZooKeeper at all.

Related solutions

AWS Architecture Review

A focused, senior-level review of your current AWS architecture, with a prioritized list of what to fix and why.

1 week
$4k–$8k
View

AWS Data Lakehouse Platform

A production data lakehouse on AWS (S3-backed data lake with Glue, Iceberg, Athena, and Redshift) structured for analytics, ML feature stores, and governed data sharing.

4–10 weeks
$25k–$60k
View

Get an event architecture that fits your traffic

We work with startups and growing companies on AWS architecture decisions. Tell us about your situation and we will recommend the right approach.

As of October 25, 2026. Provided as is, for general information, and accurate to the best of our knowledge on the date shown. Not legal, compliance, tax, financial or other professional advice. Prices, limits and vendor services change; check current vendor documentation before relying on anything here. Terms of use.