`quarkus_test_0` in sanity.external is failing (Seen on jdk17u/linux/x64)

Identified as part of https://github.com/adoptium/infrastructure/issues/4147#issuecomment-3560547543 on JDK17

Tried on four different machines and the `quarkus_test_0` fails in the same way.

Fails with [nightly](https://ci.adoptium.net/job/Grinder/15605/console) (test-ibmcloud-rhel9-x64-1) and [release](https://ci.adoptium.net/job/Grinder/15606/console) (test-azure-ubuntu2404-x64-1) builds.

[Re-run link](https://ci.adoptium.net/job/Grinder/parambuild/?SDK_RESOURCE=upstream&TARGET=quarkus_test_0&TEST_FLAG=&UPSTREAM_TEST_JOB_NAME=&DOCKER_REQUIRED=true&ACTIVE_NODE_TIMEOUT=5&VENDOR_TEST_DIRS=&EXTRA_DOCKER_ARGS=-v+%24%7BTEST_JDK_HOME%7D%3A%2Fopt%2Fjava%2Fopenjdk&TKG_OWNER_BRANCH=adoptium%3Amaster&OPENJ9_SYSTEMTEST_OWNER_BRANCH=eclipse%3Amaster&PLATFORM=x86-64_linux&GENERATE_JOBS=true&KEEP_REPORTDIR=false&PERSONAL_BUILD=false&ADOPTOPENJDK_REPO=https%3A%2F%2Fgithub.com%2Fadoptium%2Faqa-tests.git&RERUN_ITERATIONS=0&LABEL=&EXTRA_OPTIONS=&CUSTOMIZED_SDK_URL=&BUILD_IDENTIFIER=&ADOPTOPENJDK_BRANCH=master&LIGHT_WEIGHT_CHECKOUT=false&NON_AQA_TEST_REPOS=&USE_JRE=false&ARTIFACTORY_SERVER=&KEEP_WORKSPACE=false&USER_CREDENTIALS_ID=&JDK_VERSION=17&ITERATIONS=1&VENDOR_TEST_REPOS=&JDK_REPO=https%3A%2F%2Fgithub.com%2Fadoptium%2Fjdk17u&OPENJ9_BRANCH=master&OPENJ9_SHA=&JCK_GIT_REPO=&VENDOR_TEST_BRANCHES=&OPENJ9_REPO=https%3A%2F%2Fgithub.com%2Feclipse-openj9%2Fopenj9.git&UPSTREAM_JOB_NAME=build-scripts%2Fjobs%2Fjdk17u%2Fjdk17u-linux-x64-temurin&CLOUD_PROVIDER=&PLATFORM_AND_MACHINE=&CUSTOM_TARGET=&VENDOR_TEST_SHAS=&JDK_BRANCH=jdk-17.0.18%2B2_adopt&TEST_IMAGES_REQUIRED=true&LABEL_ADDITION=&ARTIFACTORY_REPO=&ARTIFACTORY_ROOT_DIR=&UPSTREAM_TEST_JOB_NUMBER=&DOCKERIMAGE_TAG=releases&JDK_IMPL=hotspot&TEST_TIME=120&SSH_AGENT_CREDENTIAL=&AUTO_DETECT=true&SLACK_CHANNEL=aqavit-bot&DYNAMIC_COMPILE=true&ADOPTOPENJDK_SYSTEMTEST_OWNER_BRANCH=adoptium%3Amaster&APPLICATION_OPTIONS=&CUSTOMIZED_SDK_URL_CREDENTIAL_ID=eclipse_temurin_bot_email_and_token&ARCHIVE_TEST_RESULTS=false&OPENJDK_SHA=&NUM_MACHINES=&RERUN_FAILURE=false&BUILD_LIST=external&USE_TESTENV_PROPERTIES=false&UPSTREAM_JOB_NUMBER=666&STF_OWNER_BRANCH=adoptium%3Amaster&TIME_LIMIT=25&JVM_OPTIONS=&PARALLEL=None)

Noting that it can take up to 2½ hours before it fails, and will have many lines with `[ERROR]` in them before getting the final failure shown below:

```
21:03:15 [INFO] BUILD FAILURE
21:03:15 [INFO] ------------------------------------------------------------------------
21:03:15 [INFO] Total time: 01:29 h
21:03:15 [INFO] Finished at: 2025-11-20T20:54:20Z
21:03:15 [INFO] ------------------------------------------------------------------------
21:03:15 [ERROR] Java heap space -> [Help 1]
21:03:15 [ERROR]
21:03:15 [ERROR] To see the full stack trace of the errors, re-run Maven with the -e switch.
21:03:15 [ERROR] Re-run Maven using the -X switch to enable full debug logging.
21:03:15 [ERROR]
21:03:15 [ERROR] For more information about the errors and possible solutions, please read the following articles:
21:03:15 [ERROR] [Help 1] http://cwiki.apache.org/confluence/display/MAVEN/OutOfMemoryError
21:03:15 -----------------------------------
21:03:15 quarkus_test_0_FAILED
21:03:15 duration_ms: 5369631
21:03:15 ]: The incoming YAML document exceeds the limit: 3145728 code points.
```

Start with the linked Grinder runs and the quarkus_test_0 target, comparing the nightly and release logs around the Maven heap failure and YAML document limit. Trace the external test setup that produces the oversized YAML or memory use. Done means the JDK 17 Linux x64 test completes successfully without the reported failure.

Written by the indexing model from the issue text.