24 Pull requests merged by 9 people

• [jobs] resolve jobs queue user on API server side

  #4897 merged Mar 6, 2025

• [API server] honor SKYPILOT_DEBUG env in server log

  #4883 merged Mar 6, 2025

• [Examples] Rename airflow DAG

  #4898 merged Mar 6, 2025

• show managed jobs user column in sky status -u

  #4889 merged Mar 5, 2025

• [Test] fixed managed job return code with --no-follow for compatibility test

  #4887 merged Mar 5, 2025

• [jobs] catch NotSupportedError for sky down --purge

  #4811 merged Mar 5, 2025

• [Dashboard] Fix Log Download

  #4844 merged Mar 4, 2025

• Serve log before termination for smoke tests

  #4691 merged Mar 4, 2025

• [Docs] Avoid back links in FAQ

  #4866 merged Mar 4, 2025

• [Storage] Fix storage deletion for all

  #4872 merged Mar 4, 2025

• [Core] Exit with non-zero code on launch/exec/logs/jobs launch/jobs logs

  #4846 merged Mar 3, 2025

• [API Server] Fix admin policy enforcement on validate and optimize

  #4820 merged Mar 3, 2025

• [Docs] New "Examples" section

  #4858 merged Mar 3, 2025

• [Docs] Add quick start to k8s getting started docs

  #4799 merged Mar 3, 2025

• [FluidStack] Add NVLINK GPUs

  #3954 merged Mar 3, 2025

• [Docs] Elevate k8s volume mounting to its own section

  #4864 merged Mar 2, 2025

• [Core] Support ARM architecture

  #4835 merged Mar 1, 2025

• [API server] hornor cgroup cpu limit when set worker numbers

  #4848 merged Mar 1, 2025

• [Existing clusters] Add --context-name arg, multi-cluster docs

  #4784 merged Mar 1, 2025

• [Utils] Quickfix serve status message

  #4854 merged Feb 28, 2025

• [APIServer] Cast job_id to int for job_status decoder

  #4832 merged Feb 28, 2025

• [API server] run API server in foreground in k8s Pod

  #4852 merged Feb 28, 2025

• Weekly smoke tests trigger on Github action

  #4833 merged Feb 28, 2025

• [Core] Only set credentials when it exists

  #4840 merged Feb 28, 2025

17 Pull requests opened by 9 people

• [Nebius] Nebius Object Storage support.

  #4838 opened Feb 27, 2025

• Support build number in nightly pypi build

  #4855 opened Feb 28, 2025

• [Docs] Update helm deployment docs

  #4865 opened Mar 2, 2025

• [Core] sky exec now waits cluster to be started

  #4867 opened Mar 3, 2025

• [Nebius] Add support Service Account credentials

  #4868 opened Mar 3, 2025

• Support cli var substitution in docker login command env

  #4871 opened Mar 4, 2025

• Fix kubernetes failure tests

  #4874 opened Mar 4, 2025

• Exclude `.pyc` and `__pycache__` files from config hash calculation to fix `test_launch_fast --kubernetes` failures

  #4880 opened Mar 5, 2025

• [API server] better error message on API version mismatch

  #4881 opened Mar 5, 2025

• [API server] use console entrypoint to start server

  #4884 opened Mar 5, 2025

• [API server] accelerate start by slowly start workers

  #4885 opened Mar 5, 2025

• [Docs] Minor updates to installation.rst

  #4888 opened Mar 5, 2025

• Updates the vast catalog to write directly to the vms.csv

  #4891 opened Mar 5, 2025

• [jobs] fix dashboard for remote API server

  #4895 opened Mar 5, 2025

• [Examples] SkyPilot + Temporal example

  #4896 opened Mar 5, 2025

• add runllm chat widget to skypilot's documentation page

  #4900 opened Mar 6, 2025

• print traceback when setting cluster to INIT

  #4901 opened Mar 6, 2025

17 Issues closed by 8 people

• [API server] API server prints debug logs when SKYPILOT_DEBUG not set

  #4882 closed Mar 6, 2025

• [Serve] Allow blue-green deployments of the controller / full cluster.

  #3703 closed Mar 6, 2025

• [Test] Backward compatibility test is broken

  #4886 closed Mar 5, 2025

• [Jobs] Fail to terminate jobs controller in abnormal state `INIT` even with `-p`

  #4626 closed Mar 5, 2025

• If a job fails it should exit with 1 instead of 0

  #4599 closed Mar 3, 2025

• [API Server] Admin policies don't work in the local API server during validate()

  #4818 closed Mar 3, 2025

• Smoke test `test_file_mounts --azure` fail on master, `rsync` command fail

  #4850 closed Mar 3, 2025

• Smoke test `test_managed_jobs_storage --azure` fail on master, Always showing `FAILED_PRECHECKS`

  #4849 closed Mar 3, 2025

• Not possible to specify multiple ports with SkyServe

  #3621 closed Mar 3, 2025

• [Serve] Expose multiple ports while using sky serve

  #3727 closed Mar 3, 2025

• [Core] Support ARM CPU on the cloud

  #4793 closed Mar 1, 2025

• [Lambda] Remove `local_ray` dependency for lambda

  #4601 closed Mar 1, 2025

• uvicorn worker numbers does not hornor cgroup resource limit

  #4847 closed Mar 1, 2025

• [Serve] Support skip failed replica in CLI

  #3547 closed Mar 1, 2025

• [k8s] Pre-flight checks for opening ports

  #3144 closed Mar 1, 2025

• [PythonAPI] SDK returns a dict with string keys for the job id for `sky.job_status`

  #4817 closed Feb 28, 2025

• [API Server] API Server restart after Pod recreation in Helm deployment

  #4771 closed Feb 28, 2025

27 Issues opened by 10 people

• [api server] forcefully killing server leaves hanging executors

  #4894 opened Mar 5, 2025

• ResourcesUnavailableError k8s error when launching on existing cluster

  #4893 opened Mar 5, 2025

• [Vast] catalog updater gets directed to stdout as opposed to the proper vms.csv file

  #4890 opened Mar 5, 2025

• [API server] Support limited compatibility between new clients and old servers

  #4879 opened Mar 5, 2025

• [API server] Refine error message of version mismatch

  #4878 opened Mar 5, 2025

• [Python API] Add `asyncio` support

  #4876 opened Mar 5, 2025

• [k8s] Better support for scale-to-zero autoscaling node pools

  #4875 opened Mar 4, 2025

• [Core] Autostop fails with `ClusterNotUpError`

  #4873 opened Mar 4, 2025

• [k8s] `sky status --k8s` does not take `kubernetes.allowed_contexts` into account.

  #4870 opened Mar 3, 2025

• [Serve] Multiple LB in the same service

  #4869 opened Mar 3, 2025

• [Image] Support custom ARM image for AWS, GCP and Azure

  #4863 opened Mar 1, 2025

• [Core] Kubernetes GPU image does not have base env activate by default, while the CPU image does

  #4862 opened Mar 1, 2025

• [Doc] document how to deploy multiple API servers and deploy server using existing ingress

  #4861 opened Mar 1, 2025

• [UX] GPU name not canonicalized when launch on kubernetes

  #4860 opened Mar 1, 2025

• [Tests] Client server API version compatibility tests

  #4859 opened Feb 28, 2025

• [API Server] HTTP API needs documentation and improvements

  #4857 opened Feb 28, 2025

• [API server] Warning when shutdown foreground server by ctrl-c

  #4856 opened Feb 28, 2025

• [Test] support run specific cases in sandbox in smoke test

  #4853 opened Feb 28, 2025

• How to pack Skypilot jobs and clusters onto GPU nodes with Kubernetes?

  #4851 opened Feb 28, 2025

• [Managed jobs] `sky jobs logs --name` does not work for completed jobs

  #4845 opened Feb 28, 2025

• Skypilot helm deployment in crash loop backoff

  #4843 opened Feb 28, 2025

• [Examples] Move pip installs to uv for faster setup

  #4842 opened Feb 27, 2025

• [Dev] Make cluster/job records a dataclass, instead of a dict

  #4841 opened Feb 27, 2025

• [Docs] Cloud authentication page should mention AWS IAM role

  #4839 opened Feb 27, 2025

• Feature request: make the action taken by skypilot-status-refresh-daemon user aware

  #4837 opened Feb 27, 2025

• Ulimit low on MacOS, but incorrect way to update it

  #4836 opened Feb 27, 2025

• [UX] sky exec CLI does not support dryrun, while sdk does

  #4834 opened Feb 27, 2025

22 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• [core] Lazy import modules to reduce importing time

  #4802 commented on Mar 6, 2025 • 22 new comments

• Introduce High Availability Service Controller

  #4564 commented on Mar 3, 2025 • 9 new comments

• Add linting for sentence case in Markdown and reST headings

  #4805 commented on Mar 5, 2025 • 3 new comments

• [SCP] Open port support

  #4490 commented on Mar 5, 2025 • 1 new comment

• Cost down smoke tests

  #4813 commented on Mar 5, 2025 • 0 new comments

• [Docs] Add docs on implementing priorities in k8s

  #4803 commented on Feb 28, 2025 • 0 new comments

• [Dev] Make the decorators type aware

  #4782 commented on Mar 4, 2025 • 0 new comments

• [UX] Auto-exclude unavailable kubernetes contexts

  #4692 commented on Mar 5, 2025 • 0 new comments

• [WIP][Serve] Enable launching multiple external LB on controller.

  #4362 commented on Mar 3, 2025 • 0 new comments

• Mitigating the Impact of Pylint's Inherent Limitations on Functionality of `format.sh`

  #4212 commented on Mar 2, 2025 • 0 new comments

• [Docs] Offline batch inference guide (static assignment)

  #4144 commented on Mar 5, 2025 • 0 new comments

• Introducing SkyPilot Guru on Gurubase.io

  #4132 commented on Mar 5, 2025 • 0 new comments

• [Serve] A patch for sync down logs

  #4036 commented on Mar 3, 2025 • 0 new comments

• `sky api start` loads Python code from the current directory instead of the pip-installed directory

  #4801 commented on Mar 5, 2025 • 0 new comments

• [Test] Unit test for helm chart

  #4786 commented on Mar 5, 2025 • 0 new comments

• [UX] Show jobs controller setup logs for the first `sky jobs launch`

  #4783 commented on Mar 5, 2025 • 0 new comments

• [Serve] Wait for fallback replicas to be ready before scaling down the old replicas.

  #3245 commented on Mar 3, 2025 • 0 new comments

• [Serve] Do not expose ports to public if replicas are in the same region as the controller

  #3720 commented on Mar 2, 2025 • 0 new comments

• [Roadmap] SkyPilot Roadmap Q1 2025

  #4760 commented on Mar 1, 2025 • 0 new comments

• Docker runtime environment is polluted with env variables from skypilot setup

  #4814 commented on Feb 27, 2025 • 0 new comments

• Feature request: Support command-line as array

  #4830 commented on Feb 27, 2025 • 0 new comments

• [Deployment] Build API server docker image from source, not nightly

  #4826 commented on Feb 27, 2025 • 0 new comments