You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 11 Next »

This page is meant as a template for writing a KIP. To create a KIP choose Tools->Copy on this page and modify with your content and replace the heading with the next KIP number and a description of your issue. Replace anything in italics with your own description.

Status

Current state: ["Under Discussion"]

Discussion thread: here

JIRA: here

Motivation

More discussions : https://github.com/apache/kafka/pull/16260#issuecomment-2159632052

Logging library used in tools module is inconsistent with core. slf4j2.x provides a cleaner and simpler approach.

By adding slf4j backends to dependencies, distributions will have their own slf4j backends, and its safe to define a slf4j provider in run-class. This allows to run kafka instance from source code with a specific provider too.

Avoid compatibility issue : The version of slf4j provider must be equal to slf4j API (https://www.slf4j.org/manual.html#compatibility). Including both in kafka distribution can avoid compatibility issue;

Public Interfaces

Update the below to upgrade slf4j, starting 4.0 version

  • dependencies.gradle
  • build.gradle

Proposed Changes

  1. Upgrade slf4j from 1.7.36 to 2.0.9+
  2. Add a new system variable to run-class (sh & bat) script to define -Dslf4j.provider. By default we use org.slf4j.reload4j.Reload4jServiceProvider
  3. Add other slf4j backend binding dependencies

Compatibility, Deprecation, and Migration Plan

Make sure users code works well when they provide their own slf4j jars

Ex : If slf4j-api is upgraded to 2.0.9, we also provide the compatible binding jars for reload4j logging in kafka itself. As we introduce this change in major version 4.0.0, this should be ok for users to upgrade their backend dependencies and we can also provide the compatible backend dependencies.

This way, users get an upgraded logging library.

Test Plan

Describe in few sentences how the KIP will be tested. We are mostly interested in system tests (since unit-tests are specific to implementation details). How will we know that the implementation works as expected? How will we know nothing broke?

Rejected Alternatives

If there are alternative ways of accomplishing the same thing, what were they? The purpose of this section is to motivate why the design is the way it is and not some other way.

  • No labels