Fix GraalVM native smoke-test OOM and configured MASS Gradle download for arm64.

[AlexeyKuznetsov-DD]

What Does This Do

• Move test_smoke_graalvm (amd64) and test_smoke_graalvm_arm64 to tier_l (20 GiB) — both build two native images concurrently and peaked at the old 16 GiB limit.
• Cap the quarkus native-image heap with quarkus.native.native-image-xmx=4g (mirrors spring's -Xmx4096M; quarkus was uncapped at ~75% of the container).
• Bound native-image parallelism on CI (-H:NumberOfThreads / -J-XX:ActiveProcessorCount, tied to KUBERNETES_CPU_REQUEST) — the container has a CPU request but no limit, so native-image otherwise sizes threads to the whole host.
• Configured the Gradle distribution download through the MASS cache for arm64.

Motivation

test_smoke_graalvm_arm64 (graalvm21/graalvm25) was failing with exit code 137 — OOMKilled. The job builds the spring-boot-3.0-native and quarkus-native images concurrently in one 16 GiB container:

Additional Notes

• Build-time/CI-only change — no production code touched.
• Validated: all six GraalVm jobs (17/21/25 × amd64/arm64) pass first-try.

[dd-octo-sts]

🟢 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	13.99 s	13.95 s	[-0.5%; +1.0%] (no difference)
startup:insecure-bank:tracing:Agent	12.96 s	13.00 s	[-1.0%; +0.4%] (no difference)
startup:petclinic:appsec:Agent	16.87 s	16.76 s	[-0.4%; +1.6%] (no difference)
startup:petclinic:iast:Agent	16.81 s	16.92 s	[-1.5%; +0.2%] (no difference)
startup:petclinic:profiling:Agent	16.32 s	16.90 s	[-7.9%; +1.0%] (no difference)
startup:petclinic:sca:Agent	16.86 s	16.74 s	[-0.1%; +1.5%] (no difference)
startup:petclinic:tracing:Agent	15.97 s	16.02 s	[-1.3%; +0.6%] (no difference)

Commit: f00ea4c5 · 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.

[PerfectSlayer]

Sounds good to me. Happy to have another pair of eye for the tier change part.

[bric3]

LGTM
Thanks for the improvement

[bric3]

note: @PerfectSlayer the tier L variables are defined here

dd-trace-java/.gitlab-ci.yml

Lines 177 to 185 in 8d8083c

	.tier_l:
	variables: &tier_l_variables
	GRADLE_WORKERS: 6
	GRADLE_MEMORY_MIN: 1G
	GRADLE_MEMORY_MAX: 6G
	KUBERNETES_CPU_REQUEST: 10
	KUBERNETES_MEMORY_REQUEST: 20Gi
	KUBERNETES_MEMORY_LIMIT: 20Gi

[sarahchen6]

TIL we configure this

[ValentinZakharov]

Looks good

[AlexeyKuznetsov-DD]

/merge

[gh-worker-devflow-routing-ef8351]

View all feedbacks in Devflow UI.

2026-06-23 17:54:43 UTC ℹ️ Start processing command /merge

---

2026-06-23 17:54:49 UTC ℹ️ MergeQueue: pull request added to the queue

The expected merge time in master is approximately 1h (p90).

---

2026-06-23 18:59:00 UTC ℹ️ MergeQueue: This merge request was merged