Quarkus Console Logging Does Not Support UTF-8 Encoding

[melodyfromheaven]

Describe the bug

I have encountered an issue with the console logging in Quarkus. It appears that the console log does not support UTF-8 encoding, which causes problems when logging messages containing non-ASCII characters, such as Korean characters.

Relate Issues
#36919
https://stackoverflow.com/questions/78465679/when-starting-quarkus-with-quarkusdev-in-intellij-idea-the-console-displays-ch

It's 2024, and it's surprising this issue still exists. Quarkus should be better than Spring, right?

Quarkus version : 3.15.1

Expected behavior

No response

Actual behavior

No response

How to Reproduce?

No response

Output of uname -a or ver

No response

Output of java -version

No response

Quarkus version or git rev

No response

Build tool (ie. output of mvnw --version or gradlew --version)

No response

Additional information

No response

[quarkus-bot]

/cc @dmlloyd (logging)

[dmlloyd]

Quarkus console logging does in fact support UTF-8 and any other character encoding supported by the JDK. The encoding used is the console encoding detected by the JDK. You can override that encoding by setting the stdout.encoding or native.encoding system property.

[geoand]

@dmlloyd so I assume there is nothing for us to do here?

[dmlloyd]

I think it should work as it is, right.

[Kitt3120]

I'd like to bring my situation into the conversation. I am currently debugging a similar issue at work.

When we updated from Quarkus 3.10.1 to 3.15.1, we noticed some weird behaviour regarding text encoding. Running our app locally works fine, but when deployed to the cluster, we get encoding issues with non UTF-8 characters. It would be fine for just our logs, but this issue also happens when using the REST client. When POSTing to an endpoint, our payload also has encoding issues.

We updated to 3.16.2 today and the problem persists.
I'll try setting the stdout.encoding or native.encoding system properties tomorrow when I am back at work.
If it doesn't fix the issue, I'll try to pin down the exact Quarkus version where this issue was introduced.

Edit:

Thanks for bringing up the stdout.encoding and file.encoding system properties. I was able to debug that our deployment was in fact using wrong encodings. We are using Red Hat's ubi8 (ubi-minimal) as our container base image. Updating from ubi-minimal 8.9-1137 to 8.10-1130 introduced the issue.

I am considering moving away from ubi8 as a fix. Quarkus micro image sounds good.

Edit 2:

We migrated to Alpine, and encoding is working now. I pinned down the issue to ubi-minimal 8.10-1088. Version 8.10-1054.1725961554 and below are fine.

I think the team should be aware of this fact. In the Quarkus docs, ubi-minimal:8.10 (and thus the newest, broken version) is recommended! See https://quarkus.io/guides/building-native-image#manually-using-the-minimal-base-image