Yandex Cloud
Search
Contact UsGet started
  • Pricing
  • Customer Stories
  • Documentation
  • Blog
  • All Services
  • System Status
    • Featured
    • Infrastructure & Network
    • Data Platform
    • Containers
    • Developer tools
    • Serverless
    • Security
    • Monitoring & Resources
    • AI for business
    • Business tools
  • All Solutions
    • By industry
    • By use case
    • Economics and Pricing
    • Security
    • Technical Support
    • Start testing with double trial credits
    • Cloud credits to scale your IT product
    • Gateway to Russia
    • Cloud for Startups
    • Center for Technologies and Society
    • Yandex Cloud Partner program
  • Pricing
  • Customer Stories
  • Documentation
  • Blog
© 2025 Direct Cursus Technology L.L.C.
Yandex Managed Service for Apache Kafka®
  • Getting started
    • Resource relationships
    • Topics and partitions
    • Brokers
    • KRaft protocol
    • Producers and consumers
    • User roles
    • Managing data schemas
    • Web interface for Apache Kafka®
    • APIs in Managed Service for Apache Kafka®
    • Host classes
    • High availability clusters
    • Networking in Managed Service for Apache Kafka®
    • Quotas and limits
    • Storage in Managed Service for Apache Kafka®
    • Connectors
    • Maintenance
    • Apache Kafka® settings
    • Apache Kafka® versioning policy
  • Access management
  • Pricing policy
  • Terraform reference
  • Yandex Monitoring metrics
  • Audit Trails events
  • Public materials
  • Release notes
  • FAQ
  1. Concepts
  2. Host classes

Apache Kafka® host classes

Written by
Yandex Cloud
Updated at December 3, 2025

The host class determines the computing power allocated for each host in a cluster. When you change the host class for a cluster, all existing hosts change to match it.

The host class determines the available disk types:

  • s2, m2, s3, m3: network-ssd, network-hdd, local-ssd, network-ssd-nonreplicated, network-ssd-io-m3.
  • b2, b3: network-ssd, network-hdd.

The available storage size does not depend on the host class. For storage limitations, see Quotas and limits.

Available host classesAvailable host classes

Hosts in Managed Service for Apache Kafka® clusters are deployed on Yandex Compute Cloud VMs. You can create these VMs on any platforms Compute Cloud supports. To learn more about the platforms, see Platforms.

The full list of possible host configurations on each platform is provided below.

Note

For cluster hosts residing in the ru-central1-d availability zone, you cannot use local SSD storage if you select Intel Cascade Lake.

Configuration types:

  • s2, s3: Standard configurations with 4:1 RAM GB to vCPU ratio.
  • m2, m3: Configurations with an increased RAM GB to vCPU ratio (8:1). These configurations may be useful for clusters with higher cache requirements.
  • b2, b3: Configurations with a guaranteed vCPU share under 100%. This host class is intended for test load, while the minimum recommended host configuration for production solutions is two vCPUs with a guaranteed share of 50%.
Host class name Number of vCPUs Guaranteed vCPU share RAM, GB Disk
size, GB
Intel Cascade Lake
b2.medium 2 50% 4 10 - 8,192
m2.micro 2 100% 16 10 - 32,736
m2.small 4 100% 32 10 - 32,736
m2.medium 6 100% 48 10 - 32,736
m2.large 8 100% 64 10 - 32,736
m2.xlarge 12 100% 96 10 - 32,736
m2.2xlarge 16 100% 128 10 - 32,736
m2.3xlarge 24 100% 192 10 - 32,736
m2.4xlarge 32 100% 256 10 - 32,736
m2.5xlarge 40 100% 320 10 - 32,736
m2.6xlarge 48 100% 384 10 - 32,736
m2.7xlarge 56 100% 448 10 - 32,736
m2.8xlarge 64 100% 512 10 - 32,736
s2.micro 2 100% 8 10 - 32,736
s2.small 4 100% 16 10 - 32,736
s2.medium 8 100% 32 10 - 32,736
s2.large 12 100% 48 10 - 32,736
s2.xlarge 16 100% 64 10 - 32,736
s2.2xlarge 24 100% 96 10 - 32,736
s2.3xlarge 32 100% 128 10 - 32,736
s2.4xlarge 40 100% 160 10 - 32,736
s2.5xlarge 48 100% 192 10 - 32,736
s2.6xlarge 64 100% 256 10 - 32,736
Intel Ice Lake
b3-c1-m4 2 50% 4 10 - 8,192
s3-c2-m8 2 100% 8 10 - 32,736
s3-c4-m16 4 100% 16 10 - 32,736
s3-c8-m32 8 100% 32 10 - 32,736
s3-c12-m48 12 100% 48 10 - 32,736
s3-c16-m64 16 100% 64 10 - 32,736
s3-c24-m96 24 100% 96 10 - 32,736
s3-c32-m128 32 100% 128 10 - 32,736
s3-c40-m160 40 100% 160 10 - 32,736
s3-c48-m192 48 100% 192 10 - 32,736
s3-c64-m256 64 100% 256 10 - 32,736
s3-c80-m320 80 100% 320 10 - 32,736
s3-c96-m576 96 100% 576 10 - 32,736
m3-c2-m16 2 100% 16 10 - 32,736
m3-c4-m32 4 100% 32 10 - 32,736
m3-c6-m48 6 100% 48 10 - 32,736
m3-c8-m64 8 100% 64 10 - 32,736
m3-c12-m96 12 100% 96 10 - 32,736
m3-c16-m128 16 100% 128 10 - 32,736
m3-c24-m192 24 100% 192 10 - 32,736
m3-c32-m256 32 100% 256 10 - 32,736
m3-c40-m320 40 100% 320 10 - 32,736
m3-c48-m384 48 100% 384 10 - 32,736
m3-c56-m448 56 100% 448 10 - 32,736
m3-c64-m512 64 100% 512 10 - 32,736
m3-c80-m640 80 100% 640 10 - 32,736
AMD Zen 4
s4a-c2-m8 2 100% 8 10 - 32,768
s4a-c4-m16 4 100% 16 10 - 32,768
s4a-c8-m32 8 100% 32 10 - 32,768
s4a-c16-m64 16 100% 64 10 - 32,768
s4a-c32-m128 32 100% 128 10 - 32,768
s4a-c64-m256 64 100% 256 10 - 32,768
s4a-c96-m384 96 100% 384 10 - 32,768
s4a-c128-m512 128 100% 512 10 - 32,768
s4a-c224-m896 224 100% 896 10 - 32,768
s4a-c256-m1024 256 100% 1 024 10 - 32,768
m4a-c2-m16 2 100% 16 10 - 32,768
m4a-c4-m32 4 100% 32 10 - 32,768
m4a-c8-m64 8 100% 64 10 - 32,768
m4a-c16-m128 16 100% 128 10 - 32,768
m4a-c32-m256 32 100% 256 10 - 32,768
m4a-c64-m512 64 100% 512 10 - 32,768
m4a-c96-m768 96 100% 768 10 - 32,768
m4a-c128-m1024 128 100% 1 024 10 - 32,768
m4a-c224-m1792 224 100% 1 792 10 - 32,768
AMD Zen 4 HighFreq
s4af-c2-m8 2 100% 8 10 - 32,768
s4af-c4-m16 4 100% 16 10 - 32,768
s4af-c8-m32 8 100% 32 10 - 32,768
s4af-c16-m64 16 100% 64 10 - 32,768
s4af-c32-m128 32 100% 128 10 - 32,768
s4af-c48-m192 48 100% 192 10 - 32,768
s4af-c80-m320 80 100% 320 10 - 32,768
m4af-c2-m16 2 100% 16 10 - 32,768
m4af-c4-m32 4 100% 32 10 - 32,768
m4af-c8-m64 8 100% 64 10 - 32,768
m4af-c16-m128 16 100% 128 10 - 32,768
m4af-c32-m256 32 100% 256 10 - 32,768
m4af-c48-m384 48 100% 384 10 - 32,768
m4af-c80-m640 80 100% 640 10 - 32,768

Was the article helpful?

Previous
APIs in Managed Service for Apache Kafka®
Next
High availability clusters
© 2025 Direct Cursus Technology L.L.C.