kubeflow/sdk

View on GitHub

Universal Python SDK to run AI workloads on Kubernetes

Stars: 148
Forks: 263
Open beginner issues: 0
Indexed issues: 110
Avg merge: 1d 2h
Merged PRs (30d): 1

Dominant language: Python
License: Apache-2.0
Last GitHub push: Aug 29, 2026
Latest indexed: Sep 18, 2026
Contributing guide: Contributing guide
Code of conduct: No code of conduct
Beginner labels: help wanted good first issue

Beginner friendly Everything

Open All

110 open issues indexed

SparkConnect Idle clean-up Open

kind/feature needs-triage

Difficulty 5/5 Over a week Newbie friendliness 35/100

kubeflow/sdk#800 · 1 comment · Opened Sep 18, 2026
bug: workflow posts "first PR" message on every PR Open

kind/bug needs-triage

Difficulty 3/5 1-2 days Newbie friendliness 38/100

kubeflow/sdk#798 · 1 comment · 1 reaction · Opened Sep 17, 2026
bug(trainer): list_jobs and get_job fail after a referenced runtime is deleted Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

kubeflow/sdk#797 · 1 comment · Opened Sep 16, 2026
bug(trainer): list_jobs omits runtime when no filter is provided Open

Difficulty 2/5 1-3 hours Newbie friendliness 78/100

kubeflow/sdk#793 · Opened Sep 10, 2026
bug(trainer): container backend reports Complete when one container status is Unknown Open

kind/bug needs-triage

Difficulty 2/5 1-3 hours Newbie friendliness 88/100

kubeflow/sdk#787 · Opened Sep 7, 2026
bug: runtime cache never activates: every list_jobs() re-fetches from GitHub per job Open

kind/bug needs-triage

Difficulty 4/5 3-5 days Newbie friendliness 70/100

kubeflow/sdk#783 · 1 comment · Opened Sep 6, 2026
fix(trainer): PodmanClientAdapter.list_containers crashes when filters is None Open

kind/bug needs-triage

kubeflow/sdk#782 · 6 comments · 1 assignee · Opened Sep 5, 2026
Evaluate replacing pyspark-connect with pyspark-client in [spark] extra Open

area/spark

kubeflow/sdk#781 · 6 comments · 2 assignees · Opened Sep 3, 2026
SparkConnect's notion of the Spark version isn't tied to the actual image Open

area/spark kind/bug

kubeflow/sdk#780 · 4 comments · 1 assignee · Opened Sep 2, 2026
BuiltinTrainer: no dataset arg is emitted for S3DatasetInitializer or DataCacheInitializer Open

kubeflow/sdk#768 · 1 comment · 2 reactions · 1 assignee · Opened Aug 21, 2026
LocalProcessBackend crashes on Windows due to bash PATH resolution conflict (WSL relay vs Git Bash) Open

kind/bug needs-triage

kubeflow/sdk#757 · 3 comments · 1 assignee · Opened Aug 11, 2026
Bugs in OSV-Scanner workflow and helper scripts, found while porting to kubeflow/mcp-server Open

kubeflow/sdk#748 · 2 comments · 1 assignee · Opened Aug 10, 2026
Fix remaining Sphinx warnings in Spark API documentation Open

Difficulty 3/5 1-2 days Newbie friendliness 78/100

kubeflow/sdk#746 · 2 comments · Opened Aug 9, 2026
TorchTune loss and dataset source raise AttributeError instead of ValueError when not an enum Open

kubeflow/sdk#741 · 6 comments · 1 assignee · Opened Aug 8, 2026
SparkClient never turns on the Spark UI or metrics for a submitted job Open

area/observability area/spark kind/feature

kubeflow/sdk#733 · 6 comments · 1 assignee · Opened Aug 7, 2026
LocalProcessBackend accepts filesystem-unsafe job names, resulting in FileNotFoundError on Windows Open

kind/bug needs-triage

kubeflow/sdk#714 · 2 comments · 1 assignee · Opened Aug 3, 2026
fix: strip trailing newline in get_default_target_namespace when reading in-cluster serviceaccount namespace Open

kubeflow/sdk#710 · 1 comment · 1 assignee · Opened Aug 2, 2026
bug(common): get_default_target_namespace leaves trailing newline when reading in-cluster service account namespace file Open

kubeflow/sdk#708 · 1 comment · 1 assignee · Opened Aug 2, 2026
LocalProcessBackend.list_jobs() returns TrainJob objects with Unknown status instead of actual job status Open

kind/bug needs-triage

Difficulty 2/5 1-3 hours Newbie friendliness 76/100

kubeflow/sdk#703 · Opened Aug 1, 2026
Move shared Spark and Trainer test helpers to kubeflow.common Open

area/spark area/trainer good first issue help wanted

kubeflow/sdk#700 · 6 comments · 1 reaction · 1 assignee · Opened Jul 31, 2026
[Enhancement] Allow Spark job and session log streaming to wait for the driver pod Open

area/spark kind/feature

kubeflow/sdk#691 · 3 comments · 1 assignee · Opened Jul 30, 2026
Add Spark metrics collection via Spark REST API Open

area/observability area/spark kind/feature

Difficulty 4/5 3-5 days Newbie friendliness 52/100

kubeflow/sdk#689 · 2 comments · Opened Jul 30, 2026
Bug: SparkClient treats retry states as terminal failures Open

kind/bug needs-triage

kubeflow/sdk#688 · 2 comments · 1 assignee · Opened Jul 30, 2026
LocalProcessBackend silently ignores initializer, breaking the train() API contract Open

kind/bug needs-triage

kubeflow/sdk#687 · 3 comments · 2 assignees · Opened Jul 30, 2026
feat(trainer): Add log streaming and tailing support to TrainerClient.get_job_logs() Open

kind/feature needs-triage

kubeflow/sdk#682 · 4 comments · 2 assignees · Opened Jul 29, 2026
feat(client): Add context manager support (`__enter__` / `__exit__`) to `TrainerClient` and `SparkClient` Open

kind/feature needs-triage

kubeflow/sdk#669 · 2 comments · 1 reaction · 1 assignee · Opened Jul 27, 2026
fix(trainer): `get_script_for_python_packages` raises `IndexError` when `pip_index_urls` is empty Open

kubeflow/sdk#665 · 1 comment · 1 assignee · Opened Jul 26, 2026
[bug] Spark KubernetesBackend ignores client_configuration, context filtering when config_file is set, and default target namespace resolution in KubernetesBackendConfig Open

area/spark kind/bug

kubeflow/sdk#663 · 3 comments · 1 assignee · Opened Jul 26, 2026
LocalProcessBackend.wait_for_job_status does not raise RuntimeError for failed TrainJobs Open

kind/bug needs-triage

Difficulty 2/5 1-3 hours Newbie friendliness 72/100

kubeflow/sdk#662 · 1 comment · Opened Jul 26, 2026
Spark log retrieval returns an extra empty line for empty or newline-terminated output Open

area/spark kind/bug

kubeflow/sdk#656 · 4 comments · 1 assignee · Opened Jul 26, 2026
[Epic] SparkClient – Production-Ready Distributed Data Processing & ML Workflows Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

kubeflow/sdk#655 · 3 comments · Opened Jul 26, 2026
LocalProcessBackend.list_jobs() returns incorrect runtime metadata for unfiltered job listings Open

kind/bug needs-triage

Difficulty 2/5 1-3 hours Newbie friendliness 72/100

kubeflow/sdk#653 · 1 comment · Opened Jul 25, 2026
feat(spark): support connect_timeout when connecting to an existing Spark Connect server via base_url Open

area/spark kind/feature

Difficulty 3/5 1-2 days Newbie friendliness 68/100

kubeflow/sdk#650 · 2 comments · Opened Jul 25, 2026
docs: add module-level docstrings to package __init__ files Open

Difficulty 3/5 1-2 days Newbie friendliness 74/100

kubeflow/sdk#644 · 1 comment · Opened Jul 24, 2026
Enhance type safety and exception messages in TrainerClient Open

kind/feature needs-triage

Difficulty 2/5 1-3 hours Newbie friendliness 72/100

kubeflow/sdk#640 · Opened Jul 23, 2026
Title: update_trainjob_status() throttle has a TOCTOU race — concurrent callers can bypass the 5s rate limit, causing thundering-herd bursts and out-of-order updates Open

kind/bug needs-triage

kubeflow/sdk#639 · 5 comments · 1 assignee · Opened Jul 23, 2026
Share embedded function execution script template between Trainer and Spark Open

area/spark

kubeflow/sdk#638 · 9 comments · 1 assignee · Opened Jul 23, 2026
KEP-107: Convert Spark SDK Python examples to Jupyter notebooks Open

area/spark

kubeflow/sdk#635 · 6 comments · 1 assignee · Opened Jul 23, 2026
bug(trainer): Zero test coverage for LocalJob execution logic in localprocess backend Open

kind/bug needs-triage

Difficulty 3/5 1-2 days Newbie friendliness 72/100

kubeflow/sdk#633 · 1 comment · Opened Jul 23, 2026
feat: adopt structlog for structured logging Open

area/observability kind/feature

kubeflow/sdk#631 · 7 comments · 1 assignee · Opened Jul 22, 2026
Dynamic LLM/RLHF Trainers: config-driven framework support in the SDK Open

area/gsoc area/trainer kind/feature

Difficulty 5/5 Over a week Newbie friendliness 35/100

kubeflow/sdk#626 · 1 comment · 1 reaction · Opened Jul 21, 2026
feat(trainer): extend local execution backends with customizable runtime options Open

kind/feature needs-triage

Difficulty 5/5 Over a week Newbie friendliness 35/100

kubeflow/sdk#624 · 1 comment · Opened Jul 21, 2026
docs: Improve Spark documentation navigation and structure to align with other Kubeflow SDK components Open

area/spark good first issue help wanted kind/feature

kubeflow/sdk#618 · 7 comments · 1 reaction · 1 assignee · Opened Jul 19, 2026
KEP-107: Remove `SparkJob` service account defaulting in Spark package Open

area/spark good first issue help wanted kind/feature

kubeflow/sdk#616 · 7 comments · 2 reactions · 1 assignee · Opened Jul 19, 2026
HuggingFaceModelInitializer and S3ModelInitializer share a mutable default ignore_patterns list across instances Open

kind/bug needs-triage

Difficulty 2/5 1-3 hours Newbie friendliness 78/100

kubeflow/sdk#612 · 3 comments · Opened Jul 19, 2026
KEP-107: Improve Spark E2E Robustness Open

area/spark area/testing good first issue help wanted kind/feature

kubeflow/sdk#611 · 9 comments · 1 assignee · Opened Jul 19, 2026
KEP-107: Align patterns and expand coverage for Spark package unit tests Open

area/spark good first issue help wanted kind/feature

kubeflow/sdk#609 · 5 comments · 1 assignee · Opened Jul 18, 2026
KEP-107: Move shared utilities for job submission into common module Open

area/optimizer area/spark area/trainer good first issue help wanted kind/feature

kubeflow/sdk#608 · 7 comments · 1 assignee · Opened Jul 18, 2026
KEP-107: Support Spark with Kubeflow SDK Open

area/spark kind/feature

Difficulty 5/5 Over a week Newbie friendliness 25/100

kubeflow/sdk#597 · 2 comments · 1 reaction · Opened Jul 16, 2026
KEP-107: Consolidate Spark `connect` API Open

area/spark kind/discussion kind/feature

Difficulty 5/5 Over a week Newbie friendliness 30/100

kubeflow/sdk#596 · 3 comments · Opened Jul 16, 2026

kubeflow/sdk

Get new issues in your inbox