ApsaraDB for ClickHouse:Community compatible edition

Version number

Description

Related documents

25.3.2.1-1.3.2

• Parameter adjustments

  • Enabled the ttl_only_drop_parts parameter.

  • Disabled the use_uncompressed_cache parameter.

• Other

  Upgraded to community 25.3 LTS v25.3.11.20-lts.

None

Version

Description

Related documents

25.3-1.3.1

• Parameter Adjustments

  ttl_only_drop_parts, query_profiler_cpu_time_period_ns, and query_profiler_real_time_period_ns use the default kernel value.

• Bug Fixes

  • Fixed the TTL merge logic for hot and cold data tiering, allowing TTL to function correctly on cold storage even when merges are disabled.

  • BackPort TTLMerge is affected by the size of max_bytes_to_merge_at_max_space_in_pool.

• Others

  Synced with the community 25.3 LTS release, v25.3.8.23-lts.

None

Version

Description

Related documents

25.3-1.3.0

• Bug fix

  Fixed the TTL merge logic for hot/cold tiering. TTL now remains effective on cold storage even when merge is disabled.

• Others

  Incorporated the latest community 25.3 LTS version, v25.3.7.194-lts.

None

Version

Description

Related documentation

25.3-1.1.2

Parameter modifications

• max_expanded_ast_elements is set to the kernel default value.

• max_insert_block_size is set to the kernel default value.

• Set merge_with_ttl_timeout to 3600 s.

None

23.8-1.4.3

Parameter modifications

• max_expanded_ast_elements is set to the kernel default value.

• max_insert_block_size is set to the kernel default value.

• Set merge_with_ttl_timeout to 3600 s.

New features

You can use the min_free_disk_ratio_to_perform_insert and min_free_disk_bytes_to_perform_insert configuration parameters to throw an exception when the free disk space is insufficient for write operations.

Bug fixes

Fixed an instance crash caused by the "Nullptr dereference in FindOriginalNodeForOutputName" error.

Nullptr dereference in FindOriginalNodeForOutputName

Version

Description

Related documents

25.3-1.1.1

Enhancements

Adds support for configuring Kafka SASL-related parameters.

Kafka

25.3-1.1.0

This release introduces Community Edition 25.3.

• New features

  • Introduces a new JSON data type that is production-ready and fully integrated with query acceleration.

    • Provides a more compact storage format than compressed files on disk.

    • Performs up to thousands of times faster than traditional JSON storage solutions like MongoDB.

    • The only JSON storage solution that fully supports dynamic JSON paths without requiring type casting.

  • Adds the estimateCompressionRatio function to evaluate the potential impact of different compression algorithms on the data compression ratio.

  • Adds support for Bloom filters for the Parquet data type.

  • Adds support for the WithProgress format for HTTP event streams.

  • Introduces a table-level MinMax index, which stores the minimum and maximum values of the index expression for each block.

  • Supports non-equi join operations by default.

  • Introduces the BFloat16 data type, which is ideal for representing vector embeddings in AI and vector search applications.

• Optimizations

  • Introduces a query condition cache to accelerate repetitive queries. This feature optimizes selective WHERE clauses that don't benefit from the primary index, a common scenario in dashboard and observability use cases.

  • Faster parallel hash join.

  • Transitive condition inference supports transitive inference on comparison chains. For example, for the clause WHERE a < b AND b < c AND c < 5, the optimizer automatically infers and uses the additional conditions (a < 5 AND b < 5) to enhance filtering efficiency.

  • Automatic JOIN reordering. The new query_plan_join_swap_table setting optimizes the selection of the build table for JOINs.

  • An optimization for JOIN expressions, enabled by the new optimize_extract_common_expressions setting, reduces the number of hash tables and allows for better predicate pushdown.

  • Table cloning. Use the CLONE AS clause to create a copy. This requires no additional storage space, and modifications do not affect each other.

  • Introduces an APPEND mode for refreshable materialized views. In this mode, new rows are appended to the table instead of replacing the entire view.

• Performance improvements

  An approximate 11% performance improvement compared to version 23.8.

• For a full list of new features, see the 2025 Changelog.

• Clickbench official version comparison test

Version number

Description

References

None

Community-compatible Edition clusters support Enterprise SSDs (ESSDs) at Performance Level 0 (PL0).

Community-compatible Edition storage pricing

Version

Description

References

23.8

new features

You can now migrate data from a self-managed ClickHouse cluster to an ApsaraDB for ClickHouse community-compatible edition cluster in the console.

Migrate data from a self-managed ClickHouse cluster to a community-compatible cluster

Version

Description

Related documents

N/A

ApsaraDB for ClickHouse now supports the Saudi Arabia (Riyadh) region.

None

Version number

Description

References

22.8-1.19.0

• The Kafka table engine now supports the kafka_auto_offset_reset parameter. This parameter lets you set the offset synchronization strategy to earliest or latest when you create a table.

• The MergeTree table engine now supports delayed mutation and throws exceptions for excessive mutations. The usage is the same as in version 23.8.

• The sort key of the text_log system table is changed to event_time_microseconds.

• Sync from Kafka

• Add settings to delay or throw in case of too many mutations

23.8-1.4.1

• The Kafka table engine now supports the kafka_auto_offset_reset parameter. This parameter lets you set the offset synchronization strategy to earliest or latest when you create a table.

• The sort key of the text_log system table is changed to event_time_microseconds.

Sync from Kafka

Version

Description

Related documentation

23.8-1.3.2

Fixed an issue where a ReplicatedMergeTree table became read-only after an unexpected restart.

None

23.8-1.3.1

• Updated to the community release v23.8.16.40.

• Enabled text_log by default. It now records logs at the Warning, Error, and Fatal levels, which can be queried from system.text_log.

• Disabled query_thread_log by default to reduce disk space usage.

• Fixed an issue where an unexpected shutdown blocked data distribution for a distributed table.

Comparing changes

Version number

Description

References

22.8-1.18.0

• Fixed the Kafka engine crash issue.

• Fixed a data distribution blockage on distributed tables caused by an unexpected shutdown.

None

Version

Description

References

23.8

• New features

  • Supports streaming data from S3 using the S3Queue table engine.

  • Supports directly importing compressed files in ZIP, TAR, and 7z formats into ApsaraDB for ClickHouse.

  • Supports TRUNCATE DATABASE.

  • Supports direct vector addition and subtraction.

  • Supports tuple concatenation.

  • Error messages now suggest corrections for database and table names.

• Performance optimization

  Query performance for Parquet, ORC, TSV, CSV, and JSON files improves by an average of 40%, with some queries running up to 4–5 times faster.

• ClickHouse Release 23.8

• S3Queue Table Engine

• 2023 Changelog

• TRUNCATE DATABASE

• ClickHouse: Release 23.8 Webinar

Version

Description

References

22.8

• Enhancements

  • Added support for exclusive Tunnel and data compression when importing data from a MaxCompute external table.

  • Enabled text_log by default to record logs with a log level of Warning, Error, or Fatal. You can query these logs in system.text_log.

  • Disabled query_thread_log by default to prevent excessive disk space usage.

• Fixed issues

  • Fixed issues with data type conversion and value comparison between DateTime64 and Datetime or Date fields.

  • Fixed an "Invalid number of rows in Chunk" error.

Import from MaxCompute

21.8

• New features

  Introduced a fallback session lifetime mechanism for ClickHouse connections to ZooKeeper to balance loads across ZooKeeper nodes.

• Fixed issues

  Fixed an OOM issue caused by index_granularity=0.

None

Feature

Description

References

New Region

ApsaraDB for ClickHouse clusters are available in the Malaysia (Kuala Lumpur) region.

• Create a cluster

• Community-compatible Edition compute node pricing

• Community-compatible Edition storage space pricing

• Community-compatible Edition backup storage pricing

Community-compatible Edition 22.8

Bug fixes

Fixed a bug where the Date function returned incorrect results.

None

Specification changes

The C4 and S4 specifications are discontinued.

None

Feature

Description

References

Disable data backup

You can disable data backup for ApsaraDB for ClickHouse clusters.

Data backup

Feature

Description

References

Scheduled restart of clusters

New feature

ApsaraDB for ClickHouse now lets you schedule cluster restarts.

Restart an ApsaraDB for ClickHouse cluster

Tag management

New feature

You can now manage your ApsaraDB for ClickHouse clusters with tags.

Manage tags

ZooKeeper configuration

New feature

ApsaraDB for ClickHouse clusters include a default ZooKeeper configuration of 4 cores and 8 GB of memory at no additional charge.

None

Feature

Description

References

Monitoring and alerting

New feature

View monitoring data in the ApsaraDB for ClickHouse console

Major engine version upgrade

New feature

You can now upgrade the major engine version of an ApsaraDB for ClickHouse Community-Compatible Edition cluster by using data migration.

Upgrade via data migration

Feature

Description

References

Release and attach CLB instances

New feature

You can now release and attach Classic Load Balancer (CLB) instances for ApsaraDB for ClickHouse Community-compatible Edition clusters created after December 1, 2021.

Manage CLB instances

Account management

New feature

You can now configure database accounts for Community-compatible Edition clusters using SQL statements.

• Manage database accounts of an ApsaraDB for ClickHouse cluster

• CREATE USER

• GRANT

Feature

Description

References

Community-compatible Edition V22.8

• New features

  • Expands the supported time range for Date32 and DateTime64. The new range is 1900-01-01 to 2299-12-31.

  • Adds support for lightweight delete.

  • Adds support for SQL user-defined functions and executable user-defined functions.

  • Adds support for partition-level preprocessing to address pre-aggregation issues.

• Performance improvements

  • Doubles the performance of INSERT operations.

  • Triples the performance of IN queries that contain long lists.

• Date32

• Datetime64

• DELETE statement

• SQL user-defined functions

• Executable user-defined functions

• Projections

Feature

Description

References

parameter configuration

new feature

For ApsaraDB for ClickHouse clusters created after December 1, 2021, you can now modify parameters in XML mode and view the change history.

Configure config.xml parameters

Feature

Description

References

Upgrade minor engine version

New feature

You can now upgrade the minor engine version for ApsaraDB for ClickHouse clusters purchased after December 1, 2021. Minor version upgrades provide performance improvements, new features, and bug fixes.

Upgrade minor version

Set a maintenance window

New feature

You can now set a maintenance window for one or more ApsaraDB for ClickHouse clusters. The service upgrades the clusters within the specified maintenance window to ensure their stability.

Set a maintenance window

Feature

Description

References

Monitoring and alerting

• New features

  New metrics include:

  • Number of inactive data parts

  • Number of TCP connections

  • Number of init queries run

  • Number of HTTP connections

  • Number of MaterializeMySQL tasks

  • Number of failed insert queries

  • Number of Kafka external tables

  • Number of failed select queries

  • Number of MaterializeMySQL synchronization failures

  • Number of consumption errors for Kafka external tables

• Enhancements

  Enhances cluster availability and exception monitoring for improved operational stability.

View monitoring data in the ApsaraDB for ClickHouse console

Feature

Description

References

Backup and recovery

New feature

ApsaraDB for ClickHouse clusters running 20.3, 20.8, and 21.8 now support backup and recovery.

• Data backup

• Data recovery

Feature

Description

References

HTTPS

New feature

Adds support for HTTPS to encrypt data during transmission.

• Enable HTTPS

• Connect over HTTPS

MaterializeMySQL engine upgrade

• New feature

  The MaterializeMySQL engine now supports dirty data in ENUM data types.

• Enhancements

  • The optimized full data synchronization process now runs without table locks, minimizing the impact on your business.

  • The DDL synchronization process has been optimized. The system now writes exceptions from this process to a system table.

None

TCP keep-alive enabled by default

New feature

TCP keep-alive is enabled by default to ensure connection reliability.

None

Feature

Description

References

MaterializedMySQL engine upgrade

New feature

The MaterializedMySQL engine now supports the Decimal type.

None

Feature

Description

References

Community-compatible Edition V21.8

• New features

  • Supports more SQL and JSON standards.

  • Adds the MaterializedPostgreSQL table engine and database engine, enabling synchronous replication of databases and tables.

  • Adds the leftPad(), rightPad(), leftPadUTF8(), and rightPadUTF8() functions.

  • Adds the bin and unbin functions.

  • Adds support for the FIRST keyword in the ADD INDEX statement to add an index to the beginning of the index list.

  • Adds support for the Map, UInt128, Int128, UInt256, and Int256 data types in the MapAdd and MapSubTract functions.

  • Supports the DISTINCT ON (columns) expression.

• Optimization

  Reduces short-query latency when reading data from multiple columns in a table.

None

Feature

Description

References

Binary log preservation

You can now enable binary log preservation for the MaterializeMySQL engine by setting the enable_binlog_reserved configuration parameter. This feature improves the stability of the MaterializeMySQL engine by preventing data synchronization failures caused by the purging of binary logs.

MaterializeMySQL engine

Feature

Description

References

Cluster restart

ApsaraDB for ClickHouse now lets you restart a cluster. You can restart a cluster if it becomes unavailable or to resolve an issue.

Restart an ApsaraDB for ClickHouse cluster

Monitoring page optimization

We optimized the monitoring page to improve performance and usability.

None

Support for ApsaraDB RDS for MySQL analytic instance

ApsaraDB for ClickHouse now supports ApsaraDB RDS for MySQL analytic instance. You can synchronize data from an analytic instance to ApsaraDB for ClickHouse for centralized data analysis.

Create an ApsaraDB for ClickHouse instance

Feature

Description

References

Data migration between ApsaraDB for ClickHouse clusters

This feature lets you migrate data between instances with the same engine version or from an earlier version to a later one.

Migrate data between ApsaraDB for ClickHouse community-compatible edition clusters

Feature

Description

References

Multiple scale-out methods

Horizontal scaling now supports a new simple scale-out method. You can choose from the following methods:

• migration-based scale-out

• simple scale-out

Scale-in and scale-out

MaterializeMySQL engine upgrade

• Improves support for the JSON, TIME, and YEAR data types.

• Fixed bugs related to the key keyword and incremental table creation.

• Added support for the following configuration parameters:

  • skip_error_count

  • alter include_tables

MaterializeMySQL

Performance optimization for horizontal scaling

Optimizations to the clickhouse-copier tool improve horizontal scaling efficiency.

Scale a cluster

Feature

Description

References

Backup and recovery

Prevents accidental data loss and improves data security.

Data backup

Feature

Description

References

Tiered storage of hot and cold data

Automatically migrates data between hot and cold tiers, reducing storage costs by over 80% without compromising performance.

Tiered storage of hot and cold data

Support for PL2 and PL3 ESSDs

You can now use PL2 and PL3 ESSDs to significantly improve I/O throughput.

None

Disk encryption

Encrypts your entire data disk at the block storage level to safeguard your data. Even if a backup is compromised, the encrypted data remains inaccessible.

Enable disk encryption

Feature

Description

References

Availability in new regions

ApsaraDB for ClickHouse is now available in the China (Hohhot) and China (Hong Kong) regions.

None

Log delivery optimization

You can deliver log data from Log Service to ApsaraDB for ClickHouse for real-time log analysis.

Synchronize data from Log Service with ApsaraDB for ClickHouse

Feature

Description

References

New regions

ApsaraDB for ClickHouse is now available in the China (Zhangjiakou), China (Chengdu), and Japan (Tokyo) regions.

None

Account and permission management

You can now create accounts, change passwords, and modify permissions on databases and data dictionaries for new and existing users in the ApsaraDB for ClickHouse console.

Manage database accounts of an ApsaraDB for ClickHouse cluster

Global parameter modification

You can now modify global parameters in the ApsaraDB for ClickHouse console. Changes take effect after the cluster restarts.

Configure config.xml parameters

Feature

Description

References

Scale-out with balanced data distribution

Ensures even data distribution when you add a cluster node, which helps prevent data skew.

None

Query management

You can now identify a slow SQL query and terminate running queries.

Query management

New 4-core, 16 GB specification

A new 4-core, 16 GB specification is now available, improving single-node processing capacity.

Billable items

Data dictionary

You can now configure a data dictionary in the console. If your business data contains constants or is divided into a fact table and a dimension table, use a data dictionary to replace the dimension table. Querying the dictionary directly for dimension data improves query efficiency.

Dictionary management

Resource queues

Prevents an instance from crashing when a single large query consumes excessive resources—a common issue in open source ClickHouse.

ClickHouse resource queues

Feature

Description

References

Integration with Quick BI

Quick BI now supports ApsaraDB for ClickHouse as a data source for building BI systems.

Quick BI

Feature

Description

References

Version 20.3

Released engine version 20.3.10.75 to fix multiple engine bugs and enhance stability.

None

Real-time data migration and synchronization from MySQL databases

Supports data migration and real-time synchronization from MySQL databases to enable real-time, interactive analysis of your business data.

Import data from ApsaraDB RDS for MySQL or PolarDB for MySQL.

Online log delivery and analysis

You can now create a delivery task in the ApsaraDB for ClickHouse console to deliver user behavior logs from Simple Log Service to ApsaraDB for ClickHouse for online multi-dimensional analysis of your log data.

Synchronize data from Simple Log Service

Feature

Description

References

Integration with Data Management Service (DMS)

ApsaraDB for ClickHouse integrates with Data Management Service (DMS). You can use DMS to manage databases and tables, run data queries, and orchestrate and schedule SQL tasks.

None

Feature

Description

References

New instance types

The C4, C24, C64, and C104 instance types are now available, joining the existing C8 instance type and supporting HTTP ports.

Billable items