Use Java 25 as the default compiler toolchain by AlexeyKuznetsov-DD · Pull Request #12305 · DataDog/dd-trace-java

AlexeyKuznetsov-DD · 2026-08-26T15:35:37Z

What Does This Do

Uses JDK 25 as the default Java and Groovy compiler toolchain, matching the Gradle daemon JDK. Java compilation whose
toolchain matches the daemon can run in-process instead of starting a compiler JVM for each older module toolchain.

The emitted bytecode version remains explicit and is independent of the compiler JDK. Normal Java compilation uses
--release <bytecodeVersion>. Code that intentionally references newer or unpublished JDK APIs while emitting older
bytecode supplies unsetReleaseFlagReason and uses -source/-target instead.

Toolchain 8, 11, or 21 is retained only where required by compiler or source compatibility, including old Scala compilers,
JDK internals, old Error Prone/NullAway versions, legacy jars, and a JDK 21 preview API removed in JDK 25. Every non-25
selection must include an explanation; Gradle configuration fails otherwise.

This also centralizes the Kotlin DSL configureCompiler bridge and upgrades the malformed AspectJ 1.8.13 dependency used
by the Zuul instrumentation to 1.8.14 so it can be read by modern javac.

Motivation

Many modules selected JDK 8, 11, 17, or 21 only to control their bytecode level. Because those toolchains differ from the
JDK 25 Gradle daemon, Gradle starts and warms additional compiler JVMs. Using JDK 25 for compilation while preserving each
module's bytecode and API compatibility removes unnecessary process overhead without raising the runtime requirement of
the produced artifacts.

Additional Notes

Local ./gradlew assemble --rerun-tasks benchmark, with the daemon on JDK 25, one warm-up followed by three measured runs,
and the full process tree sampled:

Metric	`master`	This PR	Change
Wall time, mean	126.2 s	94.9 s	-24.8%
CPU time, mean	771 s	591 s	-23.4%
Peak RSS, mean	18.5 GB	18.3 GB	-0.9% (noise)

A fresh uncached :instrumentationLatestDepTest process-tree comparison with six Gradle workers showed the reduction in
JVM creation directly:

Metric	`master`-based baseline	This PR
Forked JVMs	80	59-65
Peak concurrent JVMs	35	26-27

That is about 15-21 fewer JVM forks (19-26%) and 8-9 fewer JVMs at peak. Groovy, Scala/Zinc, tests, and genuine older-JDK
exceptions still fork as required, so this does not eliminate all worker JVMs. Peak memory was effectively unchanged in
the repeated assemble benchmark; the demonstrated improvements are fewer processes, lower CPU use, and shorter build
time.

Validation:

./gradlew assemble
focused production and test compilation for migrated and retained-toolchain modules
both Akka HTTP 10.6 Scala test compilation tasks, plus their Java 11 test suites (296 tests, 0 failures)
agent JAR comparison against master; class-file versions were unchanged and differences were limited to expected
compiler-generated metadata and synthetic representation
full GitLab pipeline passed end to end

JDK 25 is an explicit compiler policy. When the daemon toolchain changes, update this default and its audit guard together.
The benchmark used a warm dependency cache on one machine; assemble also includes non-compilation work.

Contributor Checklist

Format the title according to the contribution guidelines
Assign the type: and (comp: or inst:) labels in addition to any other useful labels
Avoid using close, fix, or any linking keywords when referencing an issue
Use solves instead, and assign the PR milestone to the issue
Update the CODEOWNERS file on source file addition, migration, or deletion
Update public documentation with any new configuration flags or behaviors
Once approved, use merge queue to merge the PR

Jira ticket: [PROJ-IDENT]

datadog-official · 2026-08-26T17:14:37Z

🎯 Code Coverage (details)
• Patch Coverage: 100.00%
• Overall Coverage: 59.09% (-0.00%)

_{This comment will be updated automatically if new data arrives.

🔗 Commit SHA: 47ba511 | Docs | View more details | Give us feedback!}

pr-commenter · 2026-08-26T17:29:33Z

Debugger benchmarks

Parameters

	Baseline	Candidate
baseline_or_candidate	baseline	candidate
ci_job_date	1789592007	1789592334
end_time	2026-09-16T20:54:57	2026-09-16T21:00:24
git_branch	master	alexeyk/use-java-25
git_commit_sha	`0e8becf`	`47ba511`
start_time	2026-09-16T20:53:28	2026-09-16T20:58:54

See matching parameters

	Baseline	Candidate
ci_job_id	2051469863	2051469863
ci_pipeline_id	138010028	138010028
cpu_model	Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz	Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz
git_commit_date	1789591290	1789591290

Summary

Found 0 performance improvements and 0 performance regressions! Performance is the same for 9 metrics, 6 unstable metrics.

See unchanged results

scenario	Δ mean agg_http_req_duration_min	Δ mean agg_http_req_duration_p50	Δ mean agg_http_req_duration_p75	Δ mean agg_http_req_duration_p99	Δ mean throughput
scenario:noprobe	unstable [-19.648µs; +59.915µs] or [-6.009%; +18.324%]	unstable [-26.743µs; +82.724µs] or [-7.512%; +23.237%]	unstable [-32.959µs; +92.630µs] or [-8.909%; +25.039%]	unstable [-290.117µs; +526.541µs] or [-25.565%; +46.399%]	same
scenario:basic	same	same	same	unstable [-110.296µs; +194.212µs] or [-11.111%; +19.564%]	unstable [-162.228op/s; +162.228op/s] or [-7.300%; +7.300%]
scenario:loop	same	same	same	same	same

Request duration reports for reports

gantt
    title reports - request duration [CI 0.99] : candidate=None, baseline=None
    dateFormat X
    axisFormat %s
section baseline
noprobe (356.007 µs) : 335, 377
.   : milestone, 356,
basic (339.82 µs) : 332, 347
.   : milestone, 340,
loop (8.127 ms) : 8057, 8196
.   : milestone, 8127,
section candidate
noprobe (383.998 µs) : 314, 454
.   : milestone, 384,
basic (344.054 µs) : 337, 351
.   : milestone, 344,
loop (8.124 ms) : 8060, 8188
.   : milestone, 8124,

baseline results

Scenario	Request median duration [CI 0.99]
noprobe	356.007 µs [335.126 µs, 376.887 µs]
basic	339.82 µs [332.32 µs, 347.319 µs]
loop	8.127 ms [8.057 ms, 8.196 ms]

candidate results

Scenario	Request median duration [CI 0.99]
noprobe	383.998 µs [313.536 µs, 454.459 µs]
basic	344.054 µs [337.128 µs, 350.979 µs]
loop	8.124 ms [8.06 ms, 8.188 ms]

dd-octo-sts · 2026-08-26T17:33:03Z

🟢 Java Benchmark SLOs — All performance SLOs passed

Suite	Status
Startup	🟢 pass

SLO thresholds are defined here based on automatically generated metrics. A warning is raised when results are within 5% of the threshold.

PR vs. master results

Scenario	Candidate	master	Δ (95% CI of mean)
startup:insecure-bank:iast:Agent	14.87 s	14.68 s	[+0.5%; +2.1%] (maybe worse)
startup:insecure-bank:tracing:Agent	13.64 s	13.73 s	[-1.5%; +0.2%] (no difference)
startup:petclinic:appsec:Agent	17.05 s	16.93 s	[-0.2%; +1.5%] (no difference)
startup:petclinic:iast:Agent	16.93 s	17.06 s	[-1.5%; +0.0%] (no difference)
startup:petclinic:profiling:Agent	16.66 s	16.55 s	[-3.7%; +5.0%] (no difference)
startup:petclinic:sca:Agent	17.02 s	16.73 s	[+0.8%; +2.6%] (maybe worse)
startup:petclinic:tracing:Agent	15.63 s	16.19 s	[-7.6%; +0.6%] (no difference)

Commit: 47ba5112 · CI Pipeline · Benchmarking Platform UI

Load and DaCapo benchmarks can be triggered manually in the GitLab pipeline. Results will appear in the Benchmarking Platform UI after completion.

pr-commenter · 2026-08-26T17:46:36Z

Kafka / producer-benchmark

Parameters

	Baseline	Candidate
baseline_or_candidate	baseline	candidate
git_branch	master	alexeyk/use-java-25
git_commit_date	1789576632	1789591290
git_commit_sha	`0e8becf`	`47ba511`

See matching parameters

	Baseline	Candidate
ci_job_date	1789592521	1789592521
ci_job_id	2051469860	2051469860
ci_pipeline_id	138010028	138010028
cpu_model	Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz	Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz
jdkVersion	11.0.31	11.0.31
jmhVersion	1.36	1.36
jvm	/usr/lib/jvm/java-11-openjdk-amd64/bin/java	/usr/lib/jvm/java-11-openjdk-amd64/bin/java
jvmArgs	-Dhttp.proxyHost=127.0.0.1 -Dhttp.proxyPort=15002 -Dhttps.proxyHost=127.0.0.1 -Dhttps.proxyPort=15002 -Dhttp.nonProxyHosts=localhost	*.localhost
kernel_version	Linux runner-zfyrx7zua-project-304-concurrent-0-8q58mkpz 6.8.0-1031-aws #33~22.04.1-Ubuntu SMP Thu Jun 26 14:22:30 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux	Linux runner-zfyrx7zua-project-304-concurrent-0-8q58mkpz 6.8.0-1031-aws #33~22.04.1-Ubuntu SMP Thu Jun 26 14:22:30 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux
vmName	OpenJDK 64-Bit Server VM	OpenJDK 64-Bit Server VM
vmVersion	11.0.31+11-post-1ubuntu1-22.04.2-Ubuntu	11.0.31+11-post-1ubuntu1-22.04.2-Ubuntu

Summary

Found 0 performance improvements and 0 performance regressions! Performance is the same for 3 metrics, 0 unstable metrics.

See unchanged results

scenario	Δ mean throughput
scenario:not-instrumented/KafkaProduceBenchmark.benchProduce	same
scenario:only-tracing-dsm-disabled-benchmarks/KafkaProduceBenchmark.benchProduce	unsure [+350.170op/s; +6943.705op/s] or [+0.201%; +3.983%]
scenario:only-tracing-dsm-enabled-benchmarks/KafkaProduceBenchmark.benchProduce	same

pr-commenter · 2026-08-26T17:59:29Z

Kafka / consumer-benchmark

Parameters

	Baseline	Candidate
baseline_or_candidate	baseline	candidate
git_branch	master	alexeyk/use-java-25
git_commit_date	1789576632	1789591290
git_commit_sha	`0e8becf`	`47ba511`

See matching parameters

	Baseline	Candidate
ci_job_date	1789592576	1789592576
ci_job_id	2051469862	2051469862
ci_pipeline_id	138010028	138010028
cpu_model	Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz	Intel(R) Xeon(R) Platinum 8175M CPU @ 2.50GHz
jdkVersion	11.0.31	11.0.31
jmhVersion	1.36	1.36
jvm	/usr/lib/jvm/java-11-openjdk-amd64/bin/java	/usr/lib/jvm/java-11-openjdk-amd64/bin/java
jvmArgs	-Dhttp.proxyHost=127.0.0.1 -Dhttp.proxyPort=15002 -Dhttps.proxyHost=127.0.0.1 -Dhttps.proxyPort=15002 -Dhttp.nonProxyHosts=localhost	*.localhost
kernel_version	Linux runner-zfyrx7zua-project-304-concurrent-0-4ye5hlle 6.8.0-1031-aws #33~22.04.1-Ubuntu SMP Thu Jun 26 14:22:30 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux	Linux runner-zfyrx7zua-project-304-concurrent-0-4ye5hlle 6.8.0-1031-aws #33~22.04.1-Ubuntu SMP Thu Jun 26 14:22:30 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux
vmName	OpenJDK 64-Bit Server VM	OpenJDK 64-Bit Server VM
vmVersion	11.0.31+11-post-1ubuntu1-22.04.2-Ubuntu	11.0.31+11-post-1ubuntu1-22.04.2-Ubuntu

Summary

Found 0 performance improvements and 0 performance regressions! Performance is the same for 3 metrics, 0 unstable metrics.

See unchanged results

scenario	Δ mean throughput
scenario:not-instrumented/KafkaConsumerBenchmark.benchConsume	same
scenario:only-tracing-dsm-disabled-benchmarks/KafkaConsumerBenchmark.benchConsume	unsure [+2367.039op/s; +14140.772op/s] or [+0.978%; +5.841%]
scenario:only-tracing-dsm-enabled-benchmarks/KafkaConsumerBenchmark.benchConsume	same

chatgpt-codex-connector · 2026-09-11T20:05:53Z

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-09-11T20:13:27.507027Z	`0d016ae`	Draft marked ready
🔒 Security Review	✅ Completed 2026-09-11T20:09:29.069677Z	`0d016ae`	Draft marked ready

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

bric3 · 2026-09-16T21:52:48Z

I'll be reviewing this tomorrow morning.

bric3 · 2026-09-18T16:46:10Z

Compared the exact CI artifacts with the repository-pinned jardiff-cli:0.2.0:

Merge-base 0e8becf72f54: pipeline 137928271
PR head 47ba51120ced: pipeline 138010028

Summary:

No classes were added or removed, and class-file versions are unchanged.
90 classes differ. Most changes are compiler metadata, constant-pool/member ordering, or synthetic lambda numbering.
The only instruction-level differences are shown below.

Focused jardiff output

--- datadog/trace/bootstrap/instrumentation/jfr/directallocation/DirectAllocationSource.class
+++ datadog/trace/bootstrap/instrumentation/jfr/directallocation/DirectAllocationSource.class
@@ -48,22 +70,9 @@
-    ICONST_3
-    ANEWARRAY datadog/trace/bootstrap/instrumentation/jfr/directallocation/DirectAllocationSource
-    DUP
-    ICONST_0
-    GETSTATIC datadog/trace/bootstrap/instrumentation/jfr/directallocation/DirectAllocationSource.ALLOCATE_DIRECT : Ldatadog/trace/bootstrap/instrumentation/jfr/directallocation/DirectAllocationSource;
-    AASTORE
-    DUP
-    ICONST_1
-    GETSTATIC datadog/trace/bootstrap/instrumentation/jfr/directallocation/DirectAllocationSource.MMAP : Ldatadog/trace/bootstrap/instrumentation/jfr/directallocation/DirectAllocationSource;
-    AASTORE
-    DUP
-    ICONST_2
-    GETSTATIC datadog/trace/bootstrap/instrumentation/jfr/directallocation/DirectAllocationSource.JNI : Ldatadog/trace/bootstrap/instrumentation/jfr/directallocation/DirectAllocationSource;
-    AASTORE
+    INVOKESTATIC datadog/trace/bootstrap/instrumentation/jfr/directallocation/DirectAllocationSource.$values ()[Ldatadog/trace/bootstrap/instrumentation/jfr/directallocation/DirectAllocationSource;

--- inst/datadog/trace/instrumentation/kafka_clients38/TextMapExtractAdapter.classdata
+++ inst/datadog/trace/instrumentation/kafka_clients38/TextMapExtractAdapter.classdata
@@ -132,9 +132,9 @@
-    INVOKEVIRTUAL java/nio/ByteBuffer.flip ()Ljava/nio/ByteBuffer;
+    INVOKEVIRTUAL java/nio/ByteBuffer.flip ()Ljava/nio/Buffer;

--- inst/datadog/trace/instrumentation/springweb6/HandleMatchAdvice.classdata
+++ inst/datadog/trace/instrumentation/springweb6/HandleMatchAdvice.classdata
@@ -53,9 +53,9 @@
-    INVOKEVIRTUAL java/lang/Object.getClass ()Ljava/lang/Class;
+    INVOKEINTERFACE jakarta/servlet/http/HttpServletRequest.getClass ()Ljava/lang/Class; (itf)

--- inst/datadog/trace/instrumentation/springweb6/SpringWebHttpServerDecorator.classdata
+++ inst/datadog/trace/instrumentation/springweb6/SpringWebHttpServerDecorator.classdata
@@ -397,9 +404,9 @@
-    INVOKEVIRTUAL java/lang/Object.getClass ()Ljava/lang/Class;
+    INVOKEINTERFACE org/springframework/web/servlet/View.getClass ()Ljava/lang/Class; (itf)

All affected classes remain Java 8 bytecode:

The generated enum initialization moved into $values() without changing behavior.
ByteBuffer.flip(): Buffer is the Java 8-compatible descriptor.
The invokeinterface getClass() form is the expected JDK-8272715 output and is already handled by Muzzle.

Spring 6 itself requires Java 17, but that does not change the agent's Java 8 bytecode requirement. I found no additional actionable issue.

bric3 · 2026-09-18T16:16:11Z

+ * @param unsetReleaseFlagReason If not blank, `--release` option will not be applied.
 */
-def configureCompiler(AbstractCompile it, int toolchainVersion, JavaVersion compatibilityVersion = null, String unsetReleaseFlagReason = null) {
+def configureCompiler(AbstractCompile it, int toolchainVersion, JavaVersion bytecodeVersion, String unsetReleaseFlagReason) {


suggestion: I would avoid passing 25 at every call site. It matches the Gradle daemon so in it's somewhat already configured.

I'm thinking about an overload or a default param, where only the bytecodeVersion is needed. Unless, exceptionally, the toolchainVersion is, and when it's not needed use the Gradle Daemon version via JavaLanguageVersion.current()?

This should keep the explicit toolchain version only for JDK 8/11/21 exceptions.

bric3 · 2026-09-18T17:04:13Z

 tasks.named("compileMain_java11Java", JavaCompile) {
-  configureCompiler(it, 11, JavaVersion.VERSION_1_8)
+  // Do not lift this offset: these JFR bootstrap classes run on Java 8 JFR backports.
+  configureCompiler(it, 25, JavaVersion.VERSION_1_8, "Uses jdk.jfr module (Java 11+) at Java 8 bytecode")


suggestion: I would avoid repeating the java version if the gradle daemon matches. I.e. make that a default (with the gradle daemon jvm), and use exceptions where needed. See the configureCompiler discussion.

AlexeyKuznetsov-DD added tag: no release notes Changes to exclude from release notes type: refactoring comp: tooling Build & Tooling tag: ai generated Largely based on code generated by an AI or LLM labels Aug 26, 2026

AlexeyKuznetsov-DD self-assigned this Aug 26, 2026

AlexeyKuznetsov-DD requested a review from bric3 August 26, 2026 15:39

AlexeyKuznetsov-DD force-pushed the alexeyk/use-java-25 branch from 9760e9c to 7fb66ad Compare August 26, 2026 16:55

Use Java 25 as the default compiler toolchain

debad44

AlexeyKuznetsov-DD force-pushed the alexeyk/use-java-25 branch from 7fb66ad to debad44 Compare August 26, 2026 17:06

Merge branch 'master' into alexeyk/use-java-25

0d016ae

AlexeyKuznetsov-DD marked this pull request as ready for review September 11, 2026 20:05

AlexeyKuznetsov-DD requested a review from a team as a code owner September 11, 2026 20:05

dougqh approved these changes Sep 15, 2026

View reviewed changes

Merge branch 'master' into alexeyk/use-java-25

47ba511

bric3 reviewed Sep 18, 2026

View reviewed changes

Navigation Menu

Use Java 25 as the default compiler toolchain - #12305

Use Java 25 as the default compiler toolchain#12305
AlexeyKuznetsov-DD wants to merge 3 commits into
masterfrom
alexeyk/use-java-25

AlexeyKuznetsov-DD commented Aug 26, 2026

Uh oh!

datadog-official Bot commented Aug 26, 2026 •

edited by datadog-datadog-prod-us1-2 Bot

Loading

Uh oh!

pr-commenter Bot commented Aug 26, 2026 •

edited

Loading

Uh oh!

dd-octo-sts Bot commented Aug 26, 2026 •

edited

Loading

Uh oh!

pr-commenter Bot commented Aug 26, 2026 •

edited

Loading

Uh oh!

pr-commenter Bot commented Aug 26, 2026 •

edited

Loading

Uh oh!

chatgpt-codex-connector Bot commented Sep 11, 2026 •

edited

Loading

Uh oh!

bric3 commented Sep 16, 2026

Uh oh!

bric3 commented Sep 18, 2026 •

edited

Loading

Uh oh!

bric3 Sep 18, 2026 •

edited

Loading

Uh oh!

bric3 Sep 18, 2026

Uh oh!

This branch has not been deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

AlexeyKuznetsov-DD commented Aug 26, 2026

What Does This Do

Motivation

Additional Notes

Contributor Checklist

Uh oh!

datadog-official Bot commented Aug 26, 2026 • edited by datadog-datadog-prod-us1-2 Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pr-commenter Bot commented Aug 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Debugger benchmarks

Parameters

Summary

Uh oh!

dd-octo-sts Bot commented Aug 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🟢 Java Benchmark SLOs — All performance SLOs passed

Uh oh!

pr-commenter Bot commented Aug 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Kafka / producer-benchmark

Parameters

Summary

Uh oh!

pr-commenter Bot commented Aug 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Kafka / consumer-benchmark

Parameters

Summary

Uh oh!

chatgpt-codex-connector Bot commented Sep 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codex Review Summary

Uh oh!

bric3 commented Sep 16, 2026

Uh oh!

bric3 commented Sep 18, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

bric3 Sep 18, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

bric3 Sep 18, 2026

Choose a reason for hiding this comment

Uh oh!

This branch has not been deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

datadog-official Bot commented Aug 26, 2026 •

edited by datadog-datadog-prod-us1-2 Bot

Loading

pr-commenter Bot commented Aug 26, 2026 •

edited

Loading

dd-octo-sts Bot commented Aug 26, 2026 •

edited

Loading

pr-commenter Bot commented Aug 26, 2026 •

edited

Loading

pr-commenter Bot commented Aug 26, 2026 •

edited

Loading

chatgpt-codex-connector Bot commented Sep 11, 2026 •

edited

Loading

bric3 commented Sep 18, 2026 •

edited

Loading

bric3 Sep 18, 2026 •

edited

Loading