Migrate XGBoost integ tests to dlc v2 by Jyothirmaikottu · Pull Request #5852 · aws/deep-learning-containers

Jyothirmaikottu · 2026-03-30T18:56:06Z

Purpose

Test Plan

Test Result

05afc2e

Toggle if you are merging into master Branch

By default, docker image builds and tests are disabled. Two ways to run builds and tests:

Using dlc_developer_config.toml
Using this PR description (currently only supported for PyTorch, TensorFlow, vllm, and base images)

How to use the helper utility for updating dlc_developer_config.toml

Assuming your remote is called origin (you can find out more with git remote -v)...

Run default builds and tests for a particular buildspec - also commits and pushes changes to remote; Example:

python src/prepare_dlc_dev_environment.py -b </path/to/buildspec.yml> -cp origin

Enable specific tests for a buildspec or set of buildspecs - also commits and pushes changes to remote; Example:

python src/prepare_dlc_dev_environment.py -b </path/to/buildspec.yml> -t sanity_tests -cp origin

Restore TOML file when ready to merge

python src/prepare_dlc_dev_environment.py -rcp origin

NOTE: If you are creating a PR for a new framework version, please ensure success of the local, standard, rc, and efa sagemaker tests by updating the dlc_developer_config.toml file:

sagemaker_remote_tests = true
sagemaker_efa_tests = true
sagemaker_rc_tests = true
sagemaker_local_tests = true

How to use PR description

Use the code block below to uncomment commands and run the PR CodeBuild jobs. There are two commands available:

# /buildspec <buildspec_path>
- e.g.: # /buildspec pytorch/training/buildspec.yml
- If this line is commented out, dlc_developer_config.toml will be used.
# /tests <test_list>
- e.g.: # /tests sanity security ec2
- If this line is commented out, it will run the default set of tests (same as the defaults in dlc_developer_config.toml): sanity, security, ec2, ecs, eks, sagemaker, sagemaker-local.

# /buildspec <buildspec_path>
# /tests <test_list>

Toggle if you are merging into main Branch

PR Checklist

[] I ran pre-commit run --all-files locally before creating this PR. (Read DEVELOPMENT.md for details).

--- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 118 X-AI-Prompt: rebase this branch to main and prepare cr

--- X-AI-Tool: Human X-AI-Prompt: prepare-cr

feat: Add XGBoost benchmark tests and release workflow Migrate XGBoost benchmark tests from Hydra (SMFrameworksXGBoost3_0-5Tests) to pytest + SageMaker SDK, and add a GitHub Actions release workflow for the XGBoost SageMaker container. Benchmark tests cover all 7 categories from the Hydra suite (40 total test cases): training objective, tree method, max depth, num round, data size, instance type, and content type/input mode. Tests run SageMaker training jobs in the CI account using XGBoostSageMakerRole, with benchmark data accessed cross-account from the amazonai-algorithms-benchmarking bucket in 900597767885 via bucket policy. The release workflow (release-sagemaker-xgboost.yml) includes build, unit test, security test, and benchmark test jobs. Release and generate-release-spec jobs are stubbed as TODOs pending release readiness. Also creates stub directories for integration and container tests under test/xgboost/ for future migration. ai-dev-branch commit IDs: - e24272d - da1ae1e The prompts used are captured in the footers of those commits. The initial prompt was: rebase this branch to main and prepare cr --- X-AI-Handle-Time-Seconds: 118 X-AI-Line-Changes: New:567, Altered:0, Deleted:0 X-Human-Line-Changes: New:0, Altered:1, Deleted:0 X-AI-Line-Changes-Kiro-cli: New:567, Altered:0, Deleted:0 X-AI-Handle-Time-Seconds-Kiro-cli: 118 X-AI-Change-Count: 1 X-Human-Change-Count: 1 X-AI-Change-Count-Kiro-cli: 1 X-CR-Amendment: true

--- X-AI-Tool: Human X-AI-Prompt: pushed the changes and can you fix them

style: Fix pre-commit formatting for XGBoost benchmark tests Apply black/ruff formatting fixes flagged by pre-commit hooks: expand compressed dicts and function call args to one-per-line, remove extra blank lines before imports, and wrap long log lines. ai-dev-branch commit IDs: - 77d13b7 The prompts used are captured in the footers of those commits. The initial prompt was: pushed the changes and can you fix them --- X-AI-Handle-Time-Seconds: 0 X-AI-Line-Changes: New:0, Altered:0, Deleted:0 X-Human-Line-Changes: New:188, Altered:62, Deleted:0 X-AI-Change-Count: 0 X-Human-Change-Count: 1 X-CR-Amendment: false

--- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 28 X-AI-Prompt: i need to modify the wf to test it to run

--- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 20 X-AI-Prompt: use SagemakerRole for now instead of XGboostSagemakerRole

fix: Add push trigger for testing and use SageMakerRole Add temporary push trigger on xgboost-release branch to test the workflow. Switch from XGBoostSageMakerRole to SageMakerRole for benchmark tests. ai-dev-branch commit IDs: - 0851de5 - ec02156 The prompts used are captured in the footers of those commits. The initial prompt was: i need to modify the wf to test it to run --- X-AI-Handle-Time-Seconds: 48 X-AI-Line-Changes: New:2, Altered:3, Deleted:0 X-Human-Line-Changes: New:0, Altered:0, Deleted:0 X-AI-Line-Changes-Kiro-cli: New:2, Altered:3, Deleted:0 X-AI-Handle-Time-Seconds-Kiro-cli: 48 X-AI-Change-Count: 2 X-Human-Change-Count: 0 X-AI-Change-Count-Kiro-cli: 2 X-CR-Amendment: false

--- X-AI-Tool: Human X-AI-Prompt: <none - new session>

--- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 28 X-AI-Prompt: File "/tmp/codebuild-66e41599-9565-4813-9d1e-6e8c713b8ad9/output/src1593/src/2f80db70_a77f_4b71_a412_f48319e5a33e/actions-runner/_work/deep-learning-containers/deep-learning-containers/.venv/lib64/python3.12/site-packages/_pytest/config/argparsing.py", line 429, in addoption raise ValueError(f"option names {conflict} already added") ValueError: option names {'--image-uri'} already added

--- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 93 X-AI-Prompt: check for red flags in the entire test logic

fix: Remove duplicate pytest options and add training job cleanup Fix duplicate --image-uri option crash by removing redundant pytest_addoption from xgboost conftest (inherited from parent). Add try/finally cleanup to stop SageMaker training jobs on failure. Use SageMakerRole instead of XGBoostSageMakerRole. ai-dev-branch commit IDs: - 3dc5695 - 7e3e008 - 4cf6d94 The prompts used are captured in the footers of those commits. The initial prompt was: ValueError: option names {'--image-uri'} already added --- X-AI-Handle-Time-Seconds: 121 X-AI-Line-Changes: New:7, Altered:3, Deleted:25 X-Human-Line-Changes: New:0, Altered:1, Deleted:0 X-AI-Line-Changes-Kiro-cli: New:7, Altered:3, Deleted:25 X-AI-Handle-Time-Seconds-Kiro-cli: 121 X-AI-Change-Count: 2 X-Human-Change-Count: 1 X-AI-Change-Count-Kiro-cli: 2 X-CR-Amendment: false

--- X-AI-Tool: Human X-AI-Prompt: <none - new session>

chore: Rename workflow to Release - XGBoost SageMaker ai-dev-branch commit IDs: - fe321a4 The prompts used are captured in the footers of those commits. The initial prompt was: prepare cr --- X-AI-Handle-Time-Seconds: 0 X-AI-Line-Changes: New:0, Altered:0, Deleted:0 X-Human-Line-Changes: New:0, Altered:1, Deleted:0 X-AI-Change-Count: 0 X-Human-Change-Count: 1 X-CR-Amendment: false

--- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 33 X-AI-Prompt: xgboost/benchmarks/test_training_content_type.py::test_content_type[recordio-protobuf-pipe] 2026-03-30 20:56:42 Starting - Starting the training job... 2026-03-30 20:56:56 Starting - Preparing the instances for training... 2026-03-30 20:57:32 Downloading - Downloading the training image... 2026-03-30 20:58:08 Training - Training image download completed. Training in progress.../miniconda3/lib/python3.10/site-packages/sagemaker_containers/_server.py:22: UserWarning: pkg_resources is deprecated as an API. See https://setuptools.pypa.io/en/latest/pkg_resources.html. The pkg_resources package is slated for removal as early as 2025-11-30. Refrain from using this package or pin to Setuptools<81. import pkg_resources [2026-03-30:20:58:30:INFO] Imported framework sagemaker_xgboost_container.training [2026-03-30:20:58:30:INFO] Failed to parse hyperparameter objective value reg:squarederror to Json. Returning the value itself [2026-03-30:20:58:30:INFO] Failed to parse hyperparameter tree_method value exact to Json. Returning the value itself [2026-03-30:20:58:30:INFO] No GPUs detected (normal if no gpus installed) [2026-03-30:20:58:30:INFO] Running XGBoost Sagemaker in algorithm mode [2026-03-30:20:58:30:INFO] Determined 0 GPU(s) available on the instance. [2026-03-30:20:58:30:INFO] Pipe path /opt/ml/input/data/train found. [2026-03-30:20:58:30:INFO] Path /opt/ml/input/data/validation does not exist! [2026-03-30:20:58:30:INFO] files path: ['/opt/ml/input/data/train'] [2026-03-30:20:58:30:ERROR] Reporting training FAILURE [2026-03-30:20:58:30:ERROR] framework error: Traceback (most recent call last): File "/miniconda3/lib/python3.10/site-packages/sagemaker_containers/_trainer.py", line 84, in train entrypoint() File "/miniconda3/lib/python3.10/site-packages/sagemaker_xgboost_container/training.py", line 102, in main train(framework.training_env()) File "/miniconda3/lib/python3.10/site-packages/sagemaker_xgboost_container/training.py", line 98, in train run_algorithm_mode() File "/miniconda3/lib/python3.10/site-packages/sagemaker_xgboost_container/training.py", line 64, in run_algorithm_mode sagemaker_train( File "/miniconda3/lib/python3.10/site-packages/sagemaker_xgboost_container/algorithm_mode/train.py", line 280, in sagemaker_train raise exc.UserError(f"No data in training channel path {train_path}") sagemaker_algorithm_toolkit.exceptions.UserError: No data in training channel path /opt/ml/input/data/train

fix: Remove unsupported Pipe mode benchmark tests Remove recordio-protobuf and parquet Pipe mode tests as XGBoost algorithm mode does not support pipe input for these formats. ai-dev-branch commit IDs: - 187f127 The prompts used are captured in the footers of those commits. The initial prompt was: sagemaker_algorithm_toolkit.exceptions.UserError: No data in training channel path /opt/ml/input/data/train --- X-AI-Handle-Time-Seconds: 33 X-AI-Line-Changes: New:0, Altered:2, Deleted:2 X-Human-Line-Changes: New:0, Altered:0, Deleted:0 X-AI-Line-Changes-Kiro-cli: New:0, Altered:2, Deleted:2 X-AI-Handle-Time-Seconds-Kiro-cli: 33 X-AI-Change-Count: 1 X-Human-Change-Count: 0 X-AI-Change-Count-Kiro-cli: 1 X-CR-Amendment: true

--- X-AI-Tool: Human X-AI-Prompt: looks like it is rebuilding some parts of docker file is there a way to avoid it

--- X-AI-Tool: Human X-AI-Prompt: how to create a new policy

--- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 18 X-AI-Prompt: remoce XGBOOST_SAGEMAKER_ROLE

chore: Remove unused XGBOOST_SAGEMAKER_ROLE constant ai-dev-branch commit IDs: - 013eaa4 - fb55891 - 7096f0c The prompts used are captured in the footers of those commits. The initial prompt was: remoce XGBOOST_SAGEMAKER_ROLE --- X-AI-Handle-Time-Seconds: 18 X-AI-Line-Changes: New:0, Altered:0, Deleted:1 X-Human-Line-Changes: New:9, Altered:4, Deleted:0 X-AI-Line-Changes-Kiro-cli: New:0, Altered:0, Deleted:1 X-AI-Handle-Time-Seconds-Kiro-cli: 18 X-AI-Change-Count: 1 X-Human-Change-Count: 2 X-AI-Change-Count-Kiro-cli: 1 X-CR-Amendment: false

--- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 34 X-AI-Prompt: modify the wf to run each job parallely under benchmark tests

feat: Parallelize benchmark tests using strategy matrix Split benchmark-test into 7 parallel jobs using strategy.matrix, one per test module. Uses fail-fast: false so failures don't cancel other modules. ai-dev-branch commit IDs: - 2631d13 The prompts used are captured in the footers of those commits. The initial prompt was: modify the wf to run each job parallely under benchmark tests --- X-AI-Handle-Time-Seconds: 34 X-AI-Line-Changes: New:11, Altered:4, Deleted:0 X-Human-Line-Changes: New:0, Altered:0, Deleted:0 X-AI-Line-Changes-Kiro-cli: New:11, Altered:4, Deleted:0 X-AI-Handle-Time-Seconds-Kiro-cli: 34 X-AI-Change-Count: 1 X-Human-Change-Count: 0 X-AI-Change-Count-Kiro-cli: 1 X-CR-Amendment: false

--- X-AI-Tool: Human X-AI-Prompt: <none - new session>

chore: Increase benchmark test timeout to 90 minutes ai-dev-branch commit IDs: - 0dd9434 The prompts used are captured in the footers of those commits. The initial prompt was: prepare cr --- X-AI-Handle-Time-Seconds: 0 X-AI-Line-Changes: New:0, Altered:0, Deleted:0 X-Human-Line-Changes: New:0, Altered:1, Deleted:0 X-AI-Change-Count: 0 X-Human-Change-Count: 1 X-CR-Amendment: false

--- X-AI-Tool: Human X-AI-Prompt: <none - new session>

--- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 32 X-AI-Prompt: diff --git a/test/xgboost/benchmarks/conftest.py b/test/xgboost/benchmarks/conftest.py index 7cab415..dd4384f 100644 --- a/test/xgboost/benchmarks/conftest.py +++ b/test/xgboost/benchmarks/conftest.py @@ -81,9 +81,7 @@ def run_training_job( ), } - LOGGER.info( - f"Starting benchmark job: {job_name} ({instance_count}x {instance_type})" - ) + LOGGER.info(f"Starting benchmark job: {job_name} ({instance_count}x {instance_type})") sm = boto3.client("sagemaker") start = time.time() try: diff --git a/test/xgboost/benchmarks/test_training_content_type.py b/test/xgboost/benchmarks/test_training_content_type.py index 71808a8..e070bd0 100644 --- a/test/xgboost/benchmarks/test_training_content_type.py +++ b/test/xgboost/benchmarks/test_training_content_type.py @@ -41,9 +41,7 @@ BASE_HP = { "parquet-file", ], ) -def test_content_type( - image_uri, role, benchmark_bucket, dataset_path, content_type, input_mode -): +def test_content_type(image_uri, role, benchmark_bucket, dataset_path, content_type, input_mode): _, duration, desc = run_training_job( image_uri=image_uri, role=role,

--- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 27 X-AI-Prompt: test training content type and trianing num round take more than 90 mins.

fix: Pre-commit formatting and increase benchmark timeout to 150min Apply pre-commit formatting fixes and increase benchmark job timeout from 90 to 150 minutes for content_type and num_round modules which use large datasets. ai-dev-branch commit IDs: - d73bcc0 - 8a6a668 - dc2e0f0 The prompts used are captured in the footers of those commits. The initial prompt was: pre-commit formatting fixes --- X-AI-Handle-Time-Seconds: 59 X-AI-Line-Changes: New:0, Altered:2, Deleted:2 X-Human-Line-Changes: New:0, Altered:1, Deleted:0 X-AI-Line-Changes-Kiro-cli: New:0, Altered:2, Deleted:2 X-AI-Handle-Time-Seconds-Kiro-cli: 59 X-AI-Change-Count: 2 X-Human-Change-Count: 1 X-AI-Change-Count-Kiro-cli: 2 X-CR-Amendment: false

* AI changes made during Kiro-cli session --- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 118 X-AI-Prompt: rebase this branch to main and prepare cr * Human changes made during kiro-cli session after prompt completion. --- X-AI-Tool: Human X-AI-Prompt: prepare-cr * Human changes made during kiro-cli session after prompt completion. --- X-AI-Tool: Human X-AI-Prompt: pushed the changes and can you fix them * AI changes made during Kiro-cli session --- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 28 X-AI-Prompt: i need to modify the wf to test it to run * AI changes made during Kiro-cli session --- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 20 X-AI-Prompt: use SagemakerRole for now instead of XGboostSagemakerRole * Human changes made during kiro-cli session after prompt completion. --- X-AI-Tool: Human X-AI-Prompt: <none - new session> * AI changes made during Kiro-cli session --- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 28 X-AI-Prompt: File "/tmp/codebuild-66e41599-9565-4813-9d1e-6e8c713b8ad9/output/src1593/src/2f80db70_a77f_4b71_a412_f48319e5a33e/actions-runner/_work/deep-learning-containers/deep-learning-containers/.venv/lib64/python3.12/site-packages/_pytest/config/argparsing.py", line 429, in addoption raise ValueError(f"option names {conflict} already added") ValueError: option names {'--image-uri'} already added * AI changes made during Kiro-cli session --- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 93 X-AI-Prompt: check for red flags in the entire test logic * Human changes made during kiro-cli session after prompt completion. --- X-AI-Tool: Human X-AI-Prompt: <none - new session> * AI changes made during Kiro-cli session --- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 33 X-AI-Prompt: xgboost/benchmarks/test_training_content_type.py::test_content_type[recordio-protobuf-pipe] 2026-03-30 20:56:42 Starting - Starting the training job... 2026-03-30 20:56:56 Starting - Preparing the instances for training... 2026-03-30 20:57:32 Downloading - Downloading the training image... 2026-03-30 20:58:08 Training - Training image download completed. Training in progress.../miniconda3/lib/python3.10/site-packages/sagemaker_containers/_server.py:22: UserWarning: pkg_resources is deprecated as an API. See https://setuptools.pypa.io/en/latest/pkg_resources.html. The pkg_resources package is slated for removal as early as 2025-11-30. Refrain from using this package or pin to Setuptools<81. import pkg_resources [2026-03-30:20:58:30:INFO] Imported framework sagemaker_xgboost_container.training [2026-03-30:20:58:30:INFO] Failed to parse hyperparameter objective value reg:squarederror to Json. Returning the value itself [2026-03-30:20:58:30:INFO] Failed to parse hyperparameter tree_method value exact to Json. Returning the value itself [2026-03-30:20:58:30:INFO] No GPUs detected (normal if no gpus installed) [2026-03-30:20:58:30:INFO] Running XGBoost Sagemaker in algorithm mode [2026-03-30:20:58:30:INFO] Determined 0 GPU(s) available on the instance. [2026-03-30:20:58:30:INFO] Pipe path /opt/ml/input/data/train found. [2026-03-30:20:58:30:INFO] Path /opt/ml/input/data/validation does not exist! [2026-03-30:20:58:30:INFO] files path: ['/opt/ml/input/data/train'] [2026-03-30:20:58:30:ERROR] Reporting training FAILURE [2026-03-30:20:58:30:ERROR] framework error: Traceback (most recent call last): File "/miniconda3/lib/python3.10/site-packages/sagemaker_containers/_trainer.py", line 84, in train entrypoint() File "/miniconda3/lib/python3.10/site-packages/sagemaker_xgboost_container/training.py", line 102, in main train(framework.training_env()) File "/miniconda3/lib/python3.10/site-packages/sagemaker_xgboost_container/training.py", line 98, in train run_algorithm_mode() File "/miniconda3/lib/python3.10/site-packages/sagemaker_xgboost_container/training.py", line 64, in run_algorithm_mode sagemaker_train( File "/miniconda3/lib/python3.10/site-packages/sagemaker_xgboost_container/algorithm_mode/train.py", line 280, in sagemaker_train raise exc.UserError(f"No data in training channel path {train_path}") sagemaker_algorithm_toolkit.exceptions.UserError: No data in training channel path /opt/ml/input/data/train * Human changes made during kiro-cli session after prompt completion. --- X-AI-Tool: Human X-AI-Prompt: looks like it is rebuilding some parts of docker file is there a way to avoid it * Human changes made during kiro-cli session after prompt completion. --- X-AI-Tool: Human X-AI-Prompt: how to create a new policy * AI changes made during Kiro-cli session --- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 18 X-AI-Prompt: remoce XGBOOST_SAGEMAKER_ROLE * AI changes made during Kiro-cli session --- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 34 X-AI-Prompt: modify the wf to run each job parallely under benchmark tests * Human changes made during kiro-cli session after prompt completion. --- X-AI-Tool: Human X-AI-Prompt: <none - new session> * Human changes made during kiro-cli session after prompt completion. --- X-AI-Tool: Human X-AI-Prompt: <none - new session> * AI changes made during Kiro-cli session --- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 32 X-AI-Prompt: diff --git a/test/xgboost/benchmarks/conftest.py b/test/xgboost/benchmarks/conftest.py index 7cab415..dd4384f 100644 --- a/test/xgboost/benchmarks/conftest.py +++ b/test/xgboost/benchmarks/conftest.py @@ -81,9 +81,7 @@ def run_training_job( ), } - LOGGER.info( - f"Starting benchmark job: {job_name} ({instance_count}x {instance_type})" - ) + LOGGER.info(f"Starting benchmark job: {job_name} ({instance_count}x {instance_type})") sm = boto3.client("sagemaker") start = time.time() try: diff --git a/test/xgboost/benchmarks/test_training_content_type.py b/test/xgboost/benchmarks/test_training_content_type.py index 71808a8..e070bd0 100644 --- a/test/xgboost/benchmarks/test_training_content_type.py +++ b/test/xgboost/benchmarks/test_training_content_type.py @@ -41,9 +41,7 @@ BASE_HP = { "parquet-file", ], ) -def test_content_type( - image_uri, role, benchmark_bucket, dataset_path, content_type, input_mode -): +def test_content_type(image_uri, role, benchmark_bucket, dataset_path, content_type, input_mode): _, duration, desc = run_training_job( image_uri=image_uri, role=role, * AI changes made during Kiro-cli session --- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 27 X-AI-Prompt: test training content type and trianing num round take more than 90 mins. * remove trigger * fix format

--- X-AI-Tool: Human X-AI-Prompt: i have the workflow setup for xgboost in https://github.com/aws/deep-learning-containers ready until release through pr - #5852 and i want to add the genereate release spec and release image steps like vllm - https://github.com/aws/deep-learning-containers/actions/runs/23257998179/job/67626225687. ALso check for local DLContainersReleaseLogicV2 repo to understand what have been added to the release logic to add xgboost config and comment out the benchmark tests https://github.com/aws/deep-learning-containers/blob/main/.github/workflows/release-sagemaker-xgboost.yml since they already passed in the pr and now i wanna test the gamma release

feat: Add generate-release-spec and release-image steps to XGBoost release workflow Wire the release pipeline for XGBoost SageMaker gamma testing by adding generate-release-spec and release-image jobs modeled after the vLLM SageMaker auto-release workflow. The generate-release-spec job gates on unit-test and security-test, checks if release is enabled in config, then generates the release spec. The release-image job calls the reusable-release-image workflow with the environment from config (gamma). Benchmark tests are commented out since they already passed in PR #5852. Also set public_registry to false since xgboost uses private ECR only. ai-dev-branch commit IDs: bf9bfc1 The prompts used are captured in the footers of those commits. The initial prompt was: do it --- X-AI-Handle-Time-Seconds: 28 X-AI-Line-Changes: New:38, Altered:38, Deleted:0 X-Human-Line-Changes: New:0, Altered:0, Deleted:0 X-AI-Line-Changes-Kiro-cli: New:38, Altered:38, Deleted:0 X-AI-Handle-Time-Seconds-Kiro-cli: 28 X-AI-Change-Count: 1 X-Human-Change-Count: 0 X-AI-Change-Count-Kiro-cli: 1 X-CR-Amendment: true

feat: Add release pipeline for XGBoost SageMaker gamma testing Add generate-release-spec and release-image jobs to the XGBoost SageMaker release workflow, modeled after the vLLM auto-release workflow. The generate-release-spec job gates on unit-test and security-test, checks if release is enabled in config, then generates the release spec. The release-image job calls the reusable-release-image workflow with the environment from config. Benchmark tests are commented out since they already passed in PR #5852. Also set public_registry to false since xgboost uses private ECR only. ai-dev-branch commit IDs: bd34e9a The prompts used are captured in the footers of those commits. The initial prompt was: no this has made a mess now the pr contains only that changes of 2 lines --- X-AI-Handle-Time-Seconds: 53 X-AI-Line-Changes: New:41, Altered:38, Deleted:0 X-Human-Line-Changes: New:0, Altered:0, Deleted:0 X-AI-Line-Changes-Kiro-cli: New:41, Altered:38, Deleted:0 X-AI-Handle-Time-Seconds-Kiro-cli: 53 X-AI-Change-Count: 1 X-Human-Change-Count: 0 X-AI-Change-Count-Kiro-cli: 1 X-CR-Amendment: false

Jyothirmaikottu added 3 commits March 30, 2026 11:48

AI changes made during Kiro-cli session

da1ae1e

--- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 118 X-AI-Prompt: rebase this branch to main and prepare cr

Human changes made during kiro-cli session after prompt completion.

e24272d

--- X-AI-Tool: Human X-AI-Prompt: prepare-cr

aws-deep-learning-containers-ci Bot added authorized Size:XL Determines the size of the PR labels Mar 30, 2026

Jyothirmaikottu and others added 25 commits March 30, 2026 12:50

Human changes made during kiro-cli session after prompt completion.

77d13b7

--- X-AI-Tool: Human X-AI-Prompt: pushed the changes and can you fix them

AI changes made during Kiro-cli session

ec02156

--- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 28 X-AI-Prompt: i need to modify the wf to test it to run

AI changes made during Kiro-cli session

0851de5

--- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 20 X-AI-Prompt: use SagemakerRole for now instead of XGboostSagemakerRole

Human changes made during kiro-cli session after prompt completion.

4cf6d94

--- X-AI-Tool: Human X-AI-Prompt: <none - new session>

AI changes made during Kiro-cli session

3dc5695

--- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 93 X-AI-Prompt: check for red flags in the entire test logic

Human changes made during kiro-cli session after prompt completion.

fe321a4

--- X-AI-Tool: Human X-AI-Prompt: <none - new session>

Human changes made during kiro-cli session after prompt completion.

7096f0c

--- X-AI-Tool: Human X-AI-Prompt: looks like it is rebuilding some parts of docker file is there a way to avoid it

Human changes made during kiro-cli session after prompt completion.

fb55891

--- X-AI-Tool: Human X-AI-Prompt: how to create a new policy

AI changes made during Kiro-cli session

013eaa4

--- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 18 X-AI-Prompt: remoce XGBOOST_SAGEMAKER_ROLE

AI changes made during Kiro-cli session

2631d13

--- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 34 X-AI-Prompt: modify the wf to run each job parallely under benchmark tests

Human changes made during kiro-cli session after prompt completion.

0dd9434

--- X-AI-Tool: Human X-AI-Prompt: <none - new session>

Merge branch 'main' into xgboost-release

5e5da7c

Human changes made during kiro-cli session after prompt completion.

dc2e0f0

--- X-AI-Tool: Human X-AI-Prompt: <none - new session>

AI changes made during Kiro-cli session

d73bcc0

--- X-AI-Tool: Kiro-cli X-AI-Handle-Time-Seconds: 27 X-AI-Prompt: test training content type and trianing num round take more than 90 mins.

Jyothirmaikottu force-pushed the xgboost-release branch from 5e5fa6c to b01aac9 Compare March 31, 2026 18:15

remove trigger

bb2e621

Jyothirmaikottu force-pushed the xgboost-release branch from b01aac9 to bb2e621 Compare March 31, 2026 18:27

Jyothirmaikottu and others added 2 commits March 31, 2026 11:27

Merge branch 'main' into xgboost-release

b68dd4c

fix format

5321042

Jyothirmaikottu enabled auto-merge (squash) March 31, 2026 18:58

bhanutejagk reviewed Mar 31, 2026

View reviewed changes

Comment thread test/xgboost/container/__init__.py

Comment thread test/xgboost/integration/__init__.py

bhanutejagk approved these changes Mar 31, 2026

View reviewed changes

Jyothirmaikottu merged commit 525ef2a into main Mar 31, 2026
11 checks passed

sirutBuasai deleted the xgboost-release branch April 6, 2026 18:58

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Migrate XGBoost integ tests to dlc v2#5852

Migrate XGBoost integ tests to dlc v2#5852
Jyothirmaikottu merged 32 commits intomainfrom
xgboost-release

Jyothirmaikottu commented Mar 30, 2026 •

edited

Loading

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

Jyothirmaikottu commented Mar 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Purpose

Test Plan

Test Result

PR Checklist

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Jyothirmaikottu commented Mar 30, 2026 •

edited

Loading