Timeline for org.apache.logging.log4j:log4j-api:3.0.0 release?

[jluehe]

org.apache.logging.log4j:log4j-api is still in 3.0.0-beta2. Are there any plans to release version 3.0.0 anytime soon? Thank you!

[ppkarwasz]

There will be no 3.0.0 release of the Log4j API. This is intentional: both Log4j Core 2 and Log4j Core 3 are designed to use the same API, meaning no code changes are required when upgrading to Log4j Core 3.

The current release, 3.0.0-beta3, of Log4j Core is fully compatible with Log4j API version 2.24.x. To ensure your project always uses the correct API version for Log4j Core 3, we recommend using the log4j-bom for dependency management.

📅 Log4j Core 3 Release Timeline

Development on the 3.x line began back in 2018, around the same time as the upcoming Jackson 3.x. Since then, the main branch has been losing ground and catching up with the 2.x branch in terms of bug fixes and new features—culminating in 3.0.0-beta3, which brings us very close to parity.

At this point, all planned 3.x features have been delivered. The remaining blocker for the official 3.0.0 release is porting over a final set of changes from 2.x.

As outlined in my March email, we are currently focused on the 2.25.0 release and do not have the bandwidth to finalize work on 3.x at this time.

🙋 How You Can Help

We welcome community contributions to help bring 3.0.0 across the finish line. Here’s how you can get involved:

🛠️ Port Changes from 2.x to 3.x

Issue #3161 lists items that were completed in 2.x but not yet ported to main. While some items (e.g., GraalVM support) are more complex, many are simple cherry-picks. Porting these will bring us closer to a final release.

🧪 Test the Beta Releases

We've released several 3.x betas, but user feedback has been limited. Testing these betas and reporting bugs—or even confirming smooth upgrades—would be incredibly valuable. We expect to publish at least one more beta before finalizing 3.0.0.

Your help could significantly accelerate the release.

Thanks for being part of the Log4j community!

[jluehe]

Yes, @ppkarwasz, I'd be more than happy to contribute a PR against the 2.x branch, in time for the 2.25.0 release! Please stay tuned!

[jluehe]

Created Issue #3686

[jluehe]

@ppkarwasz, for your review: #3690

[jluehe]

@ppkarwasz, can you tell me what is the ETA for the 2.25.0 release? Even an approximate date would help! Thanks!

[ppkarwasz]

You can follow the progress of release 2.25.0 in its dedicated mailing list thread.

[garydgregory]

And welcome to the pain and confusing we are forcing on our users with this confusing version policy 😞

[lockhartja]

@ppkarwasz Been using 3-Beta3 smoothly in a medium size application for the last ~10 months.....what channel should we be posting feedback

[ppkarwasz]

@lockhartja,

This discussion thread is the right place (it’s also mirrored on dev@logging, so the whole team will see it).

Have you encountered any issues with 3.0.0-beta3? From your experience, does it feel stable enough for a 3.0.0 release?

[lockhartja]

#@ppkarwasz missed your original question -- I have had Beta3 it our prod environment, running on over 2K server VMs (linux/windows) for nearly a year -- pretty standard workloads with the core feature sets of appenders, layouts, and plugins Console, JSONTemplate, and even Opentel -- very minimal issues to note and nothing that is a show stopper -- and all the issues I encountered have had documented workaround -- i.e. the ShadowJar and merging of service files -- bottom line, my only reason for periodically checking on the repository is to ensure I have the latest dependencies -- so my limited weighted opinion - a 3.0 release is a safe progression.

[ppkarwasz]

Thanks for the feedback!

While we have a lot of bug fixes and improvements to 2.x in our backlog, personally I will try to concentrate on the release of 3.x before we merge those improvements.

[rgoers]

@ppkarwasz What are we waiting on to do the initial GA release?

[ppkarwasz]

Hi @rgoers,

There are a couple of @jvz PRs that are waiting for an answer to the reviewer questions and merge. Otherwise, we could release 3.0.0 directly.

[vy]

@rgoers, there is a list of PRs that still need to be ported from 2.x to main, see #3161, plus Matt's PRs.

[daniel-b2c2]

Thanks for the update — really helpful context.

One thing I’ve been wondering: with ongoing work continuing on 2.x and new changes still needing to be ported over, is there a point where 2.x development is expected to wind down? From the outside it sometimes feels like 3.x is perpetually catching up with integrations and backports.

Is there a policy or expectation that changes should land in main before being accepted into 2.x, or is the current approach intentional? Just trying to understand the release model and how the team envisions balancing the two branches going forward.

There are some really useful changes present in 3.x that allow sentry to work with weird build systems like bazel without having to deal with implementing workarounds for merging Log4j2Plugins.dat for example.

[vy]

@daniel-b2c2, we – @garydgregory, @rgoers, @ppkarwasz, and myself – had a call regarding this subject yesterday and decided to proceed with 3.0.0 GA as it is, and as soon as possible. Let me elaborate on that...

One thing I’ve been wondering: with ongoing work continuing on 2.x and new changes still needing to be ported over, is there a point where 2.x development is expected to wind down? From the outside it sometimes feels like 3.x is perpetually catching up with integrations and backports.

Is there a policy or expectation that changes should land in main before being accepted into 2.x, or is the current approach intentional?

This is a very sharp and accurate observation. I can share my explanation for that inquiry:

1. @ppkarwasz and I, the only active maintainers since last couple of years, have been extremely busy with Log4j 2, day jobs, and private life. When maintenance tasks are ordered by decreasing priority, the release of Log4j 3 almost always showed at the bottom. There were and are still no incentives in the horizon to change that.
2. @ppkarwasz and I – but mostly @ppkarwasz – put great deal of effort into making the 3.x branch release ready in 2024 thanks to the STF grant. Yet it is an 8 year old branch, hardly synchronized with 2.x, if ever, and contains massive changes.
3. At this stage, we have concerns regarding the DI system in 3.x, there are several issues waiting to be ported (Synchronization with main backlog #3161), and the website content is worse than in 2.x, which is extra problematic given there are significant architectural changes. Plus, we are all extremely exhausted. I guess we will just release 3.0.0 (in a month or two?) and cross our fingers.

Yes, ideally only fixes should land on 2.x and main/3.x should become the development branch. But I guess now you understand why it isn't.

[lockhartja]

Good Evening Smart People -- hoping there was some well desired downtime taken -- as we are horning in on the 3rd month of the new year (and wow is it moving) -- just wanted to inquire on the state of affairs for v3 -- and apologies upfront is the question has been addressed in other forms -- not exactly sure on how to get to the mailing lists....best wishes....jl

[ppkarwasz]

Thanks for checking in and yes, time is definitely flying this year 🙂

At the moment, progress across the project is slower than we would like. Like many volunteer-driven projects, we are dealing with limited maintainer availability (currently only 5-6 hours per week across all active maintainers). In addition, we are experiencing significant bandwidth pressure from AI-generated security reports (see #4052), which has de facto almost stopped all development work.

Regarding Log4j Core 3: the codebase itself is largely in place, but it still needs some documentation and final porting work before it is ready for a release. The remaining work is not huge, roughly on the order of 40–80 hours, but given current constraints, it will likely take a couple more months.

We appreciate your patience and continued interest, and contributions are always welcome if you’d like to help move things forward.

[lockhartja]

Happy May — first allow me to express again my sincere thanks for the work that you folks do on this project. It is selfless and often thanks-less…but adds huge value to the community at large!! I would love to buy all of a cup of coffee for the time invested.

This a simple periodic check-in on the progress of 3.0 release - and while I have not seen any significant issues with the beta3 release, as a decent steward of the codebase that I maintain, I was hoping for a recommendation, from a crystal ball viewpoint, would be a more apt to rollback to v2 line given the staleness of the v3 line?

wishing you folks a healthy and wealthy May!!