rsareddy0329

feat: Add granular telemetry signals decorator params and error classification (#5963) * feat: Add granular telemetry signals with emit/emit_presence decorator params and error classification X-AI-Prompt: Implement granular telemetry signals for training, evaluation, and deployment X-AI-Tool: Kiro IDE * feat: Add granular telemetry signals decorator params and error classification * feat: Add granular telemetry signals decorator params and error classification * Add granular telemetry signals decorator params and error classification * Add granular telemetry signals decorator params and error classification * Add granular telemetry signals decorator params and error classification * Add granular telemetry signals decorator params and error classification * Add granular telemetry signals decorator params and error classification * feat: Add ATTR_TYPE, simplify error classification, add compute type tracking * feat: Add granular telemetry to CPTTrainer, add compute ATTR_TYPE to evaluators * refactor: Move camelCase conversion to _boto_functions, use maps for error classification * fix: Change evaluator to ATTR_EXISTS to avoid emitting customer ARNs --------- Co-authored-by: Roja Reddy Sareddy <rsareddy@amazon.com>

fix: Add pre-validation for Bedrock and MLflow role permissions (#6162) * feat(evaluate): Add pre-validation for execution role permissions Validate execution role permissions before submitting evaluation jobs to fail fast with actionable guidance instead of failing during execution. Reuses existing _simulate_denied_actions and _role_trusts_service from iam_role_resolver.py. Applies to all evaluators via BaseEvaluator. * refactor(evaluate): Use evaluation role type via iam_role_resolver Add 'evaluation' role type to IAM_POLICY_CONFIG with Bedrock and MLflow permissions. BaseEvaluator now calls resolve_and_validate_role with role_type='evaluation' which validates permissions and trust using the existing iam_role_resolver infrastructure. Removes the standalone role_permission_validator utility in favor of the existing centralized approach. * fix(evaluate): Address review feedback on role validation - Remove bedrock.amazonaws.com from evaluation trust policy. The serverless backend calls Bedrock using the role's own credentials; Bedrock does not assume the role, so trust is not required. - Update existing base_evaluator tests to expect role_type='evaluation'. - Remove stray blank line in llm_as_judge_evaluator.py. - Update test assertions to reflect trust policy change. * fix: scope model_eval role gate to LLM-as-Judge, address review comments * fix: scope InvokeModel to resource ARN to avoid false denials --------- Co-authored-by: Roja Reddy Sareddy <rsareddy@amazon.com>

fix(rlaif): accept preset reward_prompt template names (#6192) * fix: update timeout of RLVR sequence_length test * fix: add xdist_group to Nova Bedrock deployment tests * Address rlaif trainer fix * Address rlaif trainer fix --------- Co-authored-by: Roja Reddy Sareddy <rsareddy@amazon.com>

fix infer logic of agent_run_time from the attached input trainer (#6115) * Fix gpu integ test failure due to outdated MPG * Fix: fix agent_run_time inference from the attached input trainer --------- Co-authored-by: Roja Reddy Sareddy <rsareddy@amazon.com>

Fix: Telemetry INFO logging to print only once per process (#6083) * Fix role issue in mtrl integ tests * Fix role issue in mtrl integ tests * Fix: Telemetry INFO logging to print only once per process * Fix: Telemetry INFO logging to print only once per process * Fix: Telemetry INFO logging to print only once per process * Fix: Telemetry INFO logging to print only once per process --------- Co-authored-by: Roja Reddy Sareddy <rsareddy@amazon.com>

fix:unskip and add mock test setup for mtrl integ tests (#5937) * fix:unskip and mark MTRL tests as gpu intensive * Add Mock setup for mtrl integ tests * Add Mock setup for mtrl integ tests * Add Mock setup for mtrl integ tests --------- Co-authored-by: Roja Reddy Sareddy <rsareddy@amazon.com>

Fix role issue in mtrl integ tests (#6070) * Fix role issue in mtrl integ tests * Fix role issue in mtrl integ tests --------- Co-authored-by: Roja Reddy Sareddy <rsareddy@amazon.com>

feat(sagemaker-core): Add botocore-sync GitHub workflows (#6226) * feat(sagemaker-core): Add botocore-sync GitHub workflows Enable the daily botocore-sync automation for the sagemaker-core module in the monorepo: - Add scheduled sync (sagemaker-core-botocore-sync.yml), auto-approve, and auto-merge workflows. They trigger the CodeBuild projects via the repo's existing CI_AWS_ROLE_ARN OIDC role; auto-merge is gated to sagemaker-bot + botocore-sync* branch + "Daily Sync with Botocore" title, scoped to sagemaker-core/sample/**/*.json. - Remove a stray hardcoded absolute SERVICE_JSON_FILE_PATH override in data_extractor.py so codegen resolves service models via the package-relative constant in constants.py. * chore(sagemaker-core): Defer auto-approve/merge sync workflows Remove the auto-approve and auto-merge workflows for now so the daily botocore sync can be validated in isolation: the sync workflow opens a PR that is left for manual review. The approve/merge automation will be re-added in a follow-up once PR generation is verified. Both files remain in history (commit 99ac5f7b) and can be restored when needed. --------- Co-authored-by: Roja Reddy Sareddy <rsareddy@amazon.com>

Release 3.21.0 (#6211) Co-authored-by: Roja Reddy Sareddy <rsareddy@amazon.com>

Release 3.19.0 (#6173) * Release 3.19.0 * docs: expand 3.19.0 changelog with serve/train feature details Address review feedback: replace generic Nova follow-ups entries with the specific features shipped in #6051 — Model Reuse (serve) and job notifications, log streaming, and metrics visualization (train). --------- Co-authored-by: Roja Reddy Sareddy <rsareddy@amazon.com>