Skip to Main Content

Restrictions

To avoid various special exceptions caused by undefined behaviors in the production environment, AutoMQ for Kafka imposes certain constraints and specifications on parameters, quotas, and behaviors within the product scope. Users need to be mindful of not breaking these constraints during use. If any factual situation does not meet these constraints, please promptly submit a support ticket for assistance.

Parameter Restrictions

The following parameters, including naming and remarks for each resource, have restriction values that cannot be adjusted. Please set them strictly according to the specifications to avoid system processing anomalies caused by special characters or exceeding length limits. Most of the parameter limitations defined within the AutoMQ for Kafka product are derived from Apache Kafka. The constraints on some parameters specific to AutoMQ Cloud are detailed below.

Instance-Level Configuration

Parameter
Limit
Description
log.retention.ms
Message retention duration
  • Default value: 7 days
  • Range: 1 hour ~ 60 days
  • Customizable
The retention period for messages impacts storage space consumption and cost. It's recommended to assess business needs rationally. If the default setting does not meet your requirements, please Obtain Services▸ for assistance.
message.max.bytes
Maximum message size
  • Limit: 20MB
  • Non-customizable
If the current parameter limit does not meet your needs, please Obtain Services▸ for assistance.
offsets.retention.minutes
Offset retention period
  • Limit: 7 days
  • Non-customizable
fetch.max.bytes
Maximum data per fetch
  • Limit: 55 MB
  • Customization not supported
Single Partition Write Throughput Limit
  • Limit: 10 MBps
  • Customization not supported
Single partition read/write throughput is limited. It is recommended that applications distribute the read/write load across different partitions using different message keys to avoid partition hotspots. Please Obtaining Services▸ for assistance.
Single Partition Read Throughput Limit
  • Limit: 10 MBps
  • Customization not supported
auto.create.topics.enable
Auto-create Topic Switch
  • Default: true
  • Range: true/false
  • Customizable
It is recommended to disable the automatic creation of topics. Manage all topics through the control system and Admin API to avoid unmanageable topics.
num.partitions
Default number of partitions for created topics
  • Default value: 1
  • Range: 1~1024
  • Customizable
The default number of partitions for created topics is used when automatically creating topics. It is recommended to set this at a reasonable cluster level to avoid excessive consumption of partition quotas due to automatic topic creation.

Topic-Level Configuration

Parameter
Limit
Description
compression.type
  • Default: producer
  • Allowed Values:
    • Zstd
    • lz4
    • Producer
    • Uncompressed
    • Snappy
    • gzip
      • Custom modifications supported
Sets the final compression type for the Topic. This configuration supports ('gzip', 'snappy', 'lz4', 'zstd'). In addition to these options, 'uncompressed' is also supported, which means no compression. The 'producer' configuration value retains the producer's compression type.
cleanup.policy
Message Cleanup Policy
  • Default: delete
  • Allowed Values:
    • delete: Periodically cleans up the earliest messages based on retention time.
    • compact: Periodically merges messages based on the message key, retaining only the latest version.
      • Custom modifications supported
It is recommended to use the delete type for most business scenarios. The compact type should only be used in scenarios that require state retention.
retention.ms
  • Default: 7 days
  • Allowed Values: 1 hour ~ 60 days
  • Custom modifications supported
Sets the custom message retention period for the topic. This configuration is only effective for topics with the delete cleanup policy and will override the default cluster configuration.
max.message.bytes
  • Default: 1 MiB
  • Allowed Values: 0 B ~ 50 MiB
  • Custom modifications supported
Sets the maximum allowable record batch size for the current Topic. Messages that exceed this limit will be rejected.
message.timestamp.type
  • Default value: CreateTime
  • Limit values:
    • LogAppendTime
    • CreateTime
      • Can be customized
Defines whether the timestamp in the message is the message creation time or the log append time. The value should be 'create time' or 'log append time'.
retention.bytes
  • Default value: -1
  • Supports customization
Effective for topics with the delete cleanup policy, this controls the maximum retention space per partition. By default, there is no size limit, only a time limit. This limit is on a per-partition basis, so multiply it by the number of partitions to calculate the Topic-level retention space. This configuration overrides the cluster default configuration.
delete.retention.ms
  • Default value: 1 day
  • Limit values: 0 hours ~ 60 days
  • Supports customization
Sets the retention time for delete tombstones for Compact-type Topics. This configuration only applies to Compact-type Topics. It dictates that consumers must read the messages within this time to potentially obtain the last valid snapshot; otherwise, they may read incomplete data.
Topic naming
  • Character Limit: Only supports letters a-z or A-Z, digits 0-9, underscores (_), hyphens (-), and dots (.).
  • Length Limit: 1-250 characters.
  • Other Restrictions: Underscores (_) cannot be used at the beginning.
None.
Topic Remarks
  • Character Limit: Only supports Chinese characters, letters a-z or A-Z, digits 0-9, and underscores (_).
  • Length Limit: 1-256 characters.

Resource Quota Limits

Based on extensive production environment operation and maintenance experience, AutoMQ Cloud has implemented initial constraints on certain performance metrics and parameters within the product. In most cases, these default constraints will meet your needs. If you encounter special circumstances where they do not, please submit a support ticket for assistance.

Limitations
Limit Values
Explanation
Single Instance (Cluster) Computing Specifications
6vCPU ~ 20vCPU, providing the following capabilities:
  • Write Throughput: 240MBps~800MBps
  • Read Throughput: 240MBps~800MBps
Instance computing specifications represent the maximum throughput capacity for message processing by a single instance (cluster).
For larger-scale cluster requirements, please Obtaining Services▸ to apply for higher specifications.
Maximum Number of Topics per Instance
Determined by the computing specifications of the created instance. For details, refer to the specification limits Billing Instructions for BYOC▸.
For security and stability, it is recommended to split different business processes across different instances to avoid concentrating all operations in a single instance.
Maximum Number of Partitions per Instance (Cluster)
Single Instance (Cluster) Request QPS Limit

Product Behavior Restrictions

With over a decade of professional operation and maintenance experience, AutoMQ Kafka imposes appropriate restrictions on certain high-risk operations and features of the open-source product. These restrictions are mainly reflected in the following aspects. If your usage scenario is not satisfied, please promptly seek help through Obtaining Services▸.

Restricted Item
Restriction Description
Apache Kafka Version Compatibility
  • Conclusion: AutoMQ Kafka is compatible with Apache Kafka versions 0.9.x to 3.7.x.
  • Description: If your current version is earlier than the versions mentioned above, it is recommended to upgrade first. If you need an updated version, please refer to Obtaining Services▸ to contact us.
Apache Kafka Ecosystem Components
  • Conclusion: AutoMQ Kafka is 100% compatible with Apache Kafka, but the Business Edition currently does not provide managed services for related ecosystem components, including the following:
    • KStreams
    • Connector
    • KSQL
    • Schema Registry
      • Description: If you are currently using ecosystem components, you can continue to use your self-built solutions to connect to the AutoMQ Kafka cluster.