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 Sharded PostgreSQL
  • Getting started
    • Resource relationships
    • Sharding
    • Host classes
    • Sharded PostgreSQL storage
    • Quotas and limits
  • Access management
  • Pricing policy
  • Terraform reference
  • Release notes
  1. Concepts
  2. Host classes

Yandex Managed Service for Sharded PostgreSQL host classes

Written by
Yandex Cloud
Updated at August 14, 2025

The host class determines the computing power allocated for each host in a Managed Service for Sharded PostgreSQL cluster.

If a cluster uses standard sharding, changing the host class updates the properties of all previously created hosts.

If a cluster uses advanced sharding, the host class is changed separately for hosts with the ROUTER and COORDINATOR roles. Changing the configuration updates the properties of all previously created hosts.

Available host classesAvailable host classes

Hosts in Managed Service for Sharded PostgreSQL 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 this article.

The full list of possible host configurations on each platform is provided below. All configurations support two disk types: network-ssd and network-hdd.

The storage size does not depend on the host class.

Note

AMD Zen 3 is available to a limited number of users for private testing at the Preview stage and is not free of charge. To request access to the platform, contact your account manager.

Intel Broadwell is not available for clusters with hosts residing in the ru-central1-d availability zone.

Configuration types:

  • s1, s2, s3, and s4a: Standard configurations with a 4:1 RAM GB to vCPU ratio.

  • m2, m3, and m4a: Configurations with an increased RAM GB to vCPU ratio (8:1). These configurations may be useful for clusters with higher cache requirements.

  • c3 and c4a: Configurations with a reduced RAM GB to vCPU ratio (2:1). These configurations may be useful for clusters with higher processor performance requirements.

  • b1, 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%.

    Note

    Hosts with a guaranteed vCPU share under 50% cannot be used in multi-host clusters.

Host class name Number of CPUs CPU performance RAM, GB Disk
size, GB
Intel Broadwell
b1.medium 2 50% 4 10–512
s1.micro 2 100% 8 10–512
s1.small 4 100% 16 10–512
s1.medium 8 100% 32 10–512
s1.large 16 100% 64 10–512
s1.xlarge 32 100% 128 10–512
Intel Cascade Lake
b2.medium 2 50% 4 10–512
m2.micro 2 100% 16 10–512
m2.small 4 100% 32 10–512
m2.medium 6 100% 48 10–512
m2.large 8 100% 64 10–512
m2.xlarge 12 100% 96 10–512
m2.2xlarge 16 100% 128 10–512
m2.3xlarge 24 100% 192 10–512
m2.4xlarge 32 100% 256 10–512
m2.5xlarge 40 100% 320 10–512
m2.6xlarge 48 100% 384 10–512
m2.7xlarge 56 100% 448 10–512
m2.8xlarge 64 100% 512 10–512
s2.micro 2 100% 8 10–512
s2.small 4 100% 16 10–512
s2.medium 8 100% 32 10–512
s2.large 12 100% 48 10–512
s2.xlarge 16 100% 64 10–512
s2.2xlarge 24 100% 96 10–512
s2.3xlarge 32 100% 128 10–512
s2.4xlarge 40 100% 160 10–512
s2.5xlarge 48 100% 192 10–512
s2.6xlarge 64 100% 256 10–512
Intel Ice Lake
b3-c1-m4 2 50% 4 10–512
s3-c2-m8 2 100% 8 10–512
s3-c4-m16 4 100% 16 10–512
s3-c8-m32 8 100% 32 10–512
s3-c12-m48 12 100% 48 10–512
s3-c16-m64 16 100% 64 10–512
s3-c24-m96 24 100% 96 10–512
s3-c32-m128 32 100% 128 10–512
s3-c40-m160 40 100% 160 10–512
s3-c48-m192 48 100% 192 10–512
s3-c64-m256 64 100% 256 10–512
s3-c80-m320 80 100% 320 10–512
m3-c2-m16 2 100% 16 10–512
m3-c4-m32 4 100% 32 10–512
m3-c6-m48 6 100% 48 10–512
m3-c8-m64 8 100% 64 10–512
m3-c12-m96 12 100% 96 10–512
m3-c16-m128 16 100% 128 10–512
m3-c24-m192 24 100% 192 10–512
m3-c32-m256 32 100% 256 10–512
m3-c40-m320 40 100% 320 10–512
c3-c2-m4 2 100% 4 10–512
c3-c4-m8 4 100% 8 10–512
c3-c8-m16 8 100% 16 10–512
c3-c12-m24 12 100% 24 10–512
c3-c16-m32 16 100% 32 10–512
c3-c24-m48 24 100% 48 10–512
c3-c32-m64 32 100% 64 10–512
c3-c40-m80 40 100% 80 10–512
c3-c48-m96 48 100% 96 10–512
c3-c64-m128 64 100% 128 10–512
c3-c80-m160 80 100% 160 10–512
c3-c96-m192 96 100% 192 10–512
AMD Zen 3
s4a-c2-m8 2 100% 8 10–512
s4a-c4-m16 4 100% 16 10–512
s4a-c8-m32 8 100% 32 10–512
s4a-c12-m48 12 100% 48 10–512
s4a-c16-m64 16 100% 64 10–512
s4a-c24-m96 24 100% 96 10–512
s4a-c32-128 32 100% 128 10–512
s4a-c40-m160 40 100% 160 10–512
s4a-c48-192 48 100% 192 10–512
s4a-c64-m256 64 100% 256 10–512
s4a-c80-320 80 100% 320 10–512
s4a-c96-m576 96 100% 576 10–512
m4a-c2-m16 2 100% 16 10–512
m4a-c4-m32 4 100% 32 10–512
m4a-c6-m48 6 100% 48 10–512
m4a-c8-m64 8 100% 64 10–512
m4a-c12-m96 12 100% 96 10–512
m4a-c16-m128 16 100% 128 10–512
m4a-c24-m192 24 100% 192 10–512
m4a-c32-m256 32 100% 256 10–512
m4a-c40-m320 40 100% 320 10–512
m4a-c48-m384 48 100% 384 10–512
m4a-c56-448 56 100% 448 10–512
m4a-c64-m512 64 100% 512 10–512
m4a-c80-m640 80 100% 640 10–512
c4a-c2-m4 2 100% 4 10–512
c4a-c4-m8 4 100% 8 10–512
c4a-c8-m16 8 100% 16 10–512
c4a-c12-m24 12 100% 24 10–512
c4a-c16-m32 16 100% 32 10–512
c4a-c24-m48 24 100% 48 10–512
c4a-c32-m64 32 100% 64 10–512
c4a-c40-m80 40 100% 80 10–512
c4a-c48-m96 48 100% 96 10–512
c4a-c64-m128 64 100% 128 10–512
c4a-c80-m160 80 100% 160 10–512
c4a-c96-m192 96 100% 192 10–512

Was the article helpful?

Previous
Sharding
Next
Sharded PostgreSQL storage
© 2025 Direct Cursus Technology L.L.C.