This is an automated email from the ASF dual-hosted git repository.

morningman pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/doris-website.git


The following commit(s) were added to refs/heads/master by this push:
     new 8aa4aee3e99 docs: clarify Doris Kafka Connector batch tuning (#3793)
8aa4aee3e99 is described below

commit 8aa4aee3e99e2272da02713a58e89d717a417d36
Author: lsy3993 <[email protected]>
AuthorDate: Sat Oct 17 05:40:06 2026 +0800

    docs: clarify Doris Kafka Connector batch tuning (#3793)
    
    ## Summary
    
    Clarify batch tuning for Doris Kafka Connector tasks running on Kafka
    Connect or Confluent. The old FAQ treated an explicitly configured
    10,000-record example as the Connector default. The revised dev and 4.x
    English/Chinese entries distinguish those values, provide a
    50,000-record / 100 MiB example, and explain flush-time, task-memory,
    and latency tradeoffs.
    
    ## Validation
    
    - Checked the official Connector configuration and DorisWriter
    implementation; current defaults are 50,000 records and 100 MiB.
    - Checked all four dev/4.x English/Chinese pages and their Connector
    guide links.
    - Targeted static governance checks and git diff --check passed with no
    new blocking findings.
    - Independently reviewed by a second agent.
    
    ## Versions and languages
    
    - [x] dev
    - [x] 4.x
    - [x] English
    - [x] Chinese
    
    ---------
    
    Co-authored-by: Claude Opus 4.7 <[email protected]>
    Co-authored-by: morningman <[email protected]>
---
 docs/faq/load-faq.md                                          | 11 +++++++++++
 .../docusaurus-plugin-content-docs/current/faq/load-faq.md    | 11 +++++++++++
 .../version-4.x/faq/load-faq.md                               | 11 +++++++++++
 versioned_docs/version-4.x/faq/load-faq.md                    | 11 +++++++++++
 4 files changed, 44 insertions(+)

diff --git a/docs/faq/load-faq.md b/docs/faq/load-faq.md
index 517dae1392d..3e486b955b9 100644
--- a/docs/faq/load-faq.md
+++ b/docs/faq/load-faq.md
@@ -72,6 +72,17 @@ curl --location-trusted -u root:"" \
     http://127.0.0.1:8030/api/db/loadtest/_stream_load
 ```
 
+### Small batches when loading through Kafka Connect or Confluent
+
+When a Doris Kafka Connector task sends frequent, small Stream Load requests, 
check the task's buffer settings. For example, if an existing configuration 
sets `buffer.count.records=10000` and `buffer.size.bytes=5000000`, consider 
testing larger batches:
+
+```properties
+buffer.count.records=50000
+buffer.size.bytes=104857600
+```
+
+These are Doris Kafka Connector options, rather than general Confluent 
settings. They control the record-count and byte-size thresholds for a batch. A 
larger buffer can reduce the number of load requests, but requires more Kafka 
Connect task memory and can increase latency. Also review `buffer.flush.time`, 
which can flush a batch before either size threshold is reached. Check the 
options and defaults for the Connector version you use in [Doris Kafka 
Connector](../connection-integration/d [...]
+
 ## Routine Load 
 
 ### Major Bug Fixes
diff --git a/i18n/zh-CN/docusaurus-plugin-content-docs/current/faq/load-faq.md 
b/i18n/zh-CN/docusaurus-plugin-content-docs/current/faq/load-faq.md
index 8d384565cc1..64345212324 100644
--- a/i18n/zh-CN/docusaurus-plugin-content-docs/current/faq/load-faq.md
+++ b/i18n/zh-CN/docusaurus-plugin-content-docs/current/faq/load-faq.md
@@ -72,6 +72,17 @@ curl --location-trusted -u root:"" \
     http://127.0.0.1:8030/api/db/loadtest/_stream_load
 ```
 
+### 通过 Kafka Connect 或 Confluent 导入时批次过小
+
+Doris Kafka Connector 任务频繁发送较小的 Stream Load 请求时,应检查任务的缓冲配置。例如,已有配置显式设置了 
`buffer.count.records=10000` 和 `buffer.size.bytes=5000000` 时,可测试更大的批次:
+
+```properties
+buffer.count.records=50000
+buffer.size.bytes=104857600
+```
+
+这些参数属于 Doris Kafka Connector,并非 Confluent 
的通用配置,分别控制一个批次的记录数和字节数阈值。增大缓冲区可以减少导入请求数,但会增加 Kafka Connect 任务的内存需求,也可能增加延迟。同时检查 
`buffer.flush.time`,它可能在两个大小阈值到达前触发刷新。所用 Connector 版本的参数及默认值请参阅 [Doris Kafka 
Connector](../connection-integration/data-integration/doris-kafka-connector.md)。
+
 ## Routine Load 
 
 ### 较严重的 Bug 修复
diff --git 
a/i18n/zh-CN/docusaurus-plugin-content-docs/version-4.x/faq/load-faq.md 
b/i18n/zh-CN/docusaurus-plugin-content-docs/version-4.x/faq/load-faq.md
index 8d384565cc1..64345212324 100644
--- a/i18n/zh-CN/docusaurus-plugin-content-docs/version-4.x/faq/load-faq.md
+++ b/i18n/zh-CN/docusaurus-plugin-content-docs/version-4.x/faq/load-faq.md
@@ -72,6 +72,17 @@ curl --location-trusted -u root:"" \
     http://127.0.0.1:8030/api/db/loadtest/_stream_load
 ```
 
+### 通过 Kafka Connect 或 Confluent 导入时批次过小
+
+Doris Kafka Connector 任务频繁发送较小的 Stream Load 请求时,应检查任务的缓冲配置。例如,已有配置显式设置了 
`buffer.count.records=10000` 和 `buffer.size.bytes=5000000` 时,可测试更大的批次:
+
+```properties
+buffer.count.records=50000
+buffer.size.bytes=104857600
+```
+
+这些参数属于 Doris Kafka Connector,并非 Confluent 
的通用配置,分别控制一个批次的记录数和字节数阈值。增大缓冲区可以减少导入请求数,但会增加 Kafka Connect 任务的内存需求,也可能增加延迟。同时检查 
`buffer.flush.time`,它可能在两个大小阈值到达前触发刷新。所用 Connector 版本的参数及默认值请参阅 [Doris Kafka 
Connector](../connection-integration/data-integration/doris-kafka-connector.md)。
+
 ## Routine Load 
 
 ### 较严重的 Bug 修复
diff --git a/versioned_docs/version-4.x/faq/load-faq.md 
b/versioned_docs/version-4.x/faq/load-faq.md
index 517dae1392d..3e486b955b9 100644
--- a/versioned_docs/version-4.x/faq/load-faq.md
+++ b/versioned_docs/version-4.x/faq/load-faq.md
@@ -72,6 +72,17 @@ curl --location-trusted -u root:"" \
     http://127.0.0.1:8030/api/db/loadtest/_stream_load
 ```
 
+### Small batches when loading through Kafka Connect or Confluent
+
+When a Doris Kafka Connector task sends frequent, small Stream Load requests, 
check the task's buffer settings. For example, if an existing configuration 
sets `buffer.count.records=10000` and `buffer.size.bytes=5000000`, consider 
testing larger batches:
+
+```properties
+buffer.count.records=50000
+buffer.size.bytes=104857600
+```
+
+These are Doris Kafka Connector options, rather than general Confluent 
settings. They control the record-count and byte-size thresholds for a batch. A 
larger buffer can reduce the number of load requests, but requires more Kafka 
Connect task memory and can increase latency. Also review `buffer.flush.time`, 
which can flush a batch before either size threshold is reached. Check the 
options and defaults for the Connector version you use in [Doris Kafka 
Connector](../connection-integration/d [...]
+
 ## Routine Load 
 
 ### Major Bug Fixes


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to