Skip to content

parallize(): added batch-size() option #307

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
HofiOne wants to merge 1 commit into
base: master
Choose a base branch
from
Open

parallize(): added batch-size() option #307

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 17 additions & 0 deletions doc/_admin-guide/080_Log/000_Log_paths/004_Log_scheduler.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,6 +21,15 @@ When using the `parallelize()` method, the incoming load can be fully distribute

`parallelize()` has the following sub-options:

### batch-size()

| Type: | integer |
| Default: | 0 |

*Description:* Defines how many consecutive messages each input thread assigns to a single `parallelize()` worker.\
This preserves ordering for those messages on the output side and can also improve the performance of `parallelize()`.\
A value around 100 is recommended.

### workers()

| Type: | integer |
Expand All @@ -29,6 +38,12 @@ When using the `parallelize()` method, the incoming load can be fully distribute
*Description:* Specifies the number of worker threads (at least 1) that
{{ site.product.short_name }} uses to process incoming messages. Increasing the number of worker threads can significantly improve message-processing performance by distributing the load across all CPU cores; however, depending on the selected partitioning strategy, message ordering may be lost, even if all messages originate from a single sender.

{% include doc/admin-guide/options/deprecated-options.md old='partitions()' new='workers()' %}

![]({{ site.baseurl}}/assets/images/caution.png) **CAUTION:**
Setting an excessively high worker count can degrade performance due to increased CPU context switching, memory overhead, and resource contention. Start with a maximum value around half the number of available CPU cores, then increase it gradually while monitoring system and {{ site.product.short_name }} performance. Values significantly higher than the CPU core count rarely provide additional benefits and may even reduce overall throughput.
{: .notice--warning}

### worker-partition-key()

| Type: | template |
Expand All @@ -37,3 +52,5 @@ When using the `parallelize()` method, the incoming load can be fully distribute
*Description:* Specifies a template used for partitioning. Messages that expand the template to the same value are assigned to the same partition.

For more details, see Parallelizing message processing.

{% include doc/admin-guide/options/deprecated-options.md old='partition_key()' new='worker-partition-key()' %}