All Products
Search
Document Center

ApsaraMQ for Kafka:Instance types

Last Updated:Jan 22, 2026

This topic describes the instance types supported by ApsaraMQ for Kafka. This information helps you select the instance type that best suits your business needs.

Instance types

Note
  • Single-zone deployment: The service and data are deployed in a single zone. A zone-level failure can cause service unavailability and data loss. If you use a single-zone instance, you can create another instance in a different region and use the connector feature to back up messages. For more information, see Best practices for single-zone disaster recovery.

  • Multi-zone deployment: The service and data are deployed across multiple zones. This deployment protects against service unavailability and data loss that can result from a single-zone failure. This deployment is recommended for critical business services.

ApsaraMQ for Kafka instances are available in the following types:

Item

Standard Edition

(High-Write Version)

Professional Edition

(High-Write Version)

Professional Edition

(High-Read Version)

Serverless

Basic Edition

Serverless

Standard Edition

Serverless

Professional Edition

Storage cost

Reserved mode. Storage costs are comparable to a self-managed cluster.

If you purchase a 300 GB disk, 100 GB is for your data and 200 GB is for backups.

Reserved mode. Reduces costs by 66% compared to self-managed clusters.

If you purchase a 300 GB disk, you get 300 GB for your data and an additional 600 GB of free backup capacity.

Pay-as-you-go mode. You are billed for the storage space and duration that you use. This saves more than 70% on storage costs compared to using cloud disks for a self-managed cluster.

Compute cost

Reserved mode

Reserved mode

Pay-as-you-go mode

Service interruption

When you scale out a cluster, the system can quickly handle traffic bursts without data replication.

Elastic architecture

After a scale-out is triggered, new read and write operations support elasticity in seconds.

The separation of storage and compute enables read and write elasticity in seconds and partition migration in seconds.

Lossless and seamless vertical scaling

Not supported

Not supported

Not supported

Manually upgrade or downgrade configurations for elasticity.

Supports lossless elastic throughput up to 2x the reserved capacity.

If the reserved capacity is small, the elastic throughput limit for production and consumption is 1024 MB/s. The formula is: Max(1024 MB/s, Reserved throughput × 2)

Note

Max(a, b) returns the greater of a and b.

Peak read/write ratio

1:1

1:1

3:1

3:1

Deployment architecture

Shared instance (logically isolated)

Dedicated instance (exclusive physical cluster)

Shared instance (logically isolated)

Dedicated instance (exclusive physical cluster)

Topic TTL

Not supported

Supported for local storage

Fully supported

Message retention period

Up to 7 days

Customizable based on your business scenario

Supports unlimited storage. By default, messages can be stored for up to one year. To request a longer retention period, submit a ticket.

Disaster recovery

Compute and storage nodes are deployed in a single zone.

Instances support multi-zone deployment. If you select single-zone deployment, the compute and storage nodes are deployed in a single zone.

Compute and storage nodes are deployed in a single zone.

Multi-zone deployment across three zones (3AZ)

Performance tuning

Not supported

Customizable based on your business scenario

Customizable based on your business scenario

ACL feature

Not supported

Support

Support

Secure Sockets Layer (SSL) encryption in a virtual private cloud (VPC)

Not supported

Supported

Support

Cross-zone

Not supported

Support

Not supported

Supported

Compatible client versions

Supports client versions 0.11 to 3.x.

Supports client versions 0.11 to 3.x.

Supports client versions 0.11 to 3.x.

SLA

99.95%

Service availability is 99.99% for cross-zone deployments and 99.95% for single-zone deployments.

Provides a 99.9% Service-Level Agreement (SLA). This is slightly lower than the Serverless Standard and Professional Editions because the cluster uses more low-cost resources, such as HDDs, OSS, and Spot Instances. This edition is recommended for testing or for services with stable traffic.

For high-stability, critical services, use the Serverless Standard or Professional Edition.

Provides a high SLA of 99.95%. Recommended for production environments.

Provides a high SLA of 99.99% with three-zone disaster recovery. Instances with small reserved capacity have higher elasticity. This is the recommended enterprise-grade edition.

Disk encryption

Supported

Supported

Not supported

Number of partitions

The number of partitions depends on the instance specification:

Serverless instances

Specification Type

Number of partition replicas

Maximum number of partition replicas

Basic Edition

  • If the reserved production throughput is 1 GB/s or less, the cluster has 3,000 partition replicas.

  • If the reserved production throughput is more than 1 GB/s, 300 partition replicas are added for each 100 MB/s increase.

  • To prevent instability from excessive message fragmentation, do not create too many partition replicas for a single topic. A single topic supports a maximum of 600 partition replicas.

  • Maximum number of partition replicas per cluster: 30,000

  • Maximum number of partition replicas per topic: 600

To request a higher limit for partition replicas, please submit a ticket.

Standard Edition

Professional Edition

Subscription and pay-as-you-go instances

Standard Edition (High-Write Version)

Throughput specification

Included partitions

Maximum partitions

alikafka.hw.2xlarge

1000

4000

alikafka.hw.3xlarge

1000

4200

alikafka.hw.6xlarge

1000

4400

alikafka.hw.9xlarge

1000

4600

alikafka.hw.12xlarge

1000

4800

Professional Edition (High-Write Version)

Throughput specification

Included partitions

Maximum partitions

alikafka.hw.2xlarge

1000

4000

alikafka.hw.3xlarge

1000

4200

alikafka.hw.6xlarge

1000

4400

alikafka.hw.9xlarge

1000

4600

alikafka.hw.12xlarge

1000

4800

alikafka.hw.16xlarge

2000

5000

alikafka.hw.20xlarge

2000

6000

alikafka.hw.25xlarge

2000

7000

alikafka.hw.30xlarge

2000

8000

alikafka.hw.60xlarge

2000

9000

alikafka.hw.80xlarge

2000

10000

alikafka.hw.100xlarge

3000

12000

alikafka.hw.120xlarge

3000

14000

alikafka.hw.150xlarge

3000

16000

alikafka.hw.180xlarge

3000

18000

alikafka.hw.200xlarge

3000

20000

alikafka.hw2.220xlarge

4000

24000

alikafka.hw2.300xlarge

4000

26000

alikafka.hw2.400xlarge

4000

28000

alikafka.hw2.500xlarge

4000

30000

alikafka.hw2.600xlarge

5000

32000

alikafka.hw2.700xlarge

5000

34000

alikafka.hw2.800xlarge

5000

36000

alikafka.hw2.900xlarge

5000

38000

alikafka.hw2.1000xlarge

5000

40000

Professional Edition (High-Read Version)

Throughput specification

Included partitions

Maximum partitions

alikafka.hr.2xlarge

1000

4000

alikafka.hr.3xlarge

1000

4200

alikafka.hr.6xlarge

1000

4400

alikafka.hr.9xlarge

1000

4600

alikafka.hr.12xlarge

1000

4800

alikafka.hr.16xlarge

2000

5000

alikafka.hr.20xlarge

2000

6000

alikafka.hr.25xlarge

2000

7000

alikafka.hr.30xlarge

2000

8000

alikafka.hr.60xlarge

2000

9000

alikafka.hr.80xlarge

2000

10000

alikafka.hr.100xlarge

3000

12000

alikafka.hr.120xlarge

3000

14000

alikafka.hr.150xlarge

3000

16000

alikafka.hr.180xlarge

3000

18000

alikafka.hr.200xlarge

3000

20000

alikafka.hr2.220xlarge

4000

24000

alikafka.hr2.300xlarge

4000

26000

alikafka.hr2.400xlarge

4000

28000

alikafka.hr2.500xlarge

4000

30000

alikafka.hr2.600xlarge

5000

32000

alikafka.hr2.700xlarge

5000

34000

alikafka.hr2.800xlarge

5000

36000

alikafka.hr2.900xlarge

5000

38000

alikafka.hr2.1000xlarge

5000

40000