All pull requests

Pull requests

• [Test] Isolate test_cluster_read_permission from leaked config env

  #11070·DanielZhangQD opened Oct 14, 2026Collaborator·2 of 5 tasks·

  ·

• [Core] Ignore client overrides of admin_policy and logs

  #11069·DanielZhangQD opened Oct 14, 2026Collaborator·2 of 5 tasks·

  ·
  1

• [SSH Node Pools] Deploy hosts that need a sudo password, without exposing it

  #11067·rohansonecha opened Oct 14, 2026Collaborator·3 of 4 tasks·

  ·
  9

• [Serve] Re-resolve the upload blob on every replica launch in consolidation mode

  #11066·cg505 opened Oct 14, 2026Collaborator·

  ·

• [Provision] Install only the container packages that are missing

  #11065·cbirkhold opened Oct 13, 2026·2 of 5 tasks·

  ·
  1

  3

• [Slurm] Pick the skylet keeper's node from squeue, like the provisioner

  #11063·ryxli opened Oct 13, 2026Contributor·3 tasks·

  ·

• [API Server] Refuse requests that carry no identity under external proxy auth (hardening, opt-in)

  #11062·cbirkhold opened Oct 13, 2026·6 of 7 tasks·

  ·
  1

  12

• [Slurm] Fail fast when the login node refuses the submit account, instead of reporting a possible leak

  #11059·cg505 opened Oct 13, 2026Collaborator·7 tasks done·

  ·

• [Storage] Add Akave S3-compatible object storage support (akave://)

  #11058·Abhay-2811 opened Oct 13, 2026·2 tasks done·

  ·
  7

• [DB] Treat asyncpg shutdown, resource and read-only errors as transient

  #11057·cg505 opened Oct 13, 2026Collaborator·3 tasks done·

  ·
  2

• [Jobs] Print a finished job's logs when the followed tail never saw it RUNNING

  #11056·zpoint opened Oct 13, 2026Collaborator·

  ·
  18

• [Core] Rename workspaces while they have active resources

  #11055·DanielZhangQD opened Oct 13, 2026Collaborator·

  ·
  10

• [Slurm] Drop login shell banners from command output and allow non-GPU gres

  #11054·eugeneteoh opened Oct 13, 2026·2 of 5 tasks·

  ·
  3

• wip: kai scheduler

  #11053·ryxli opened Oct 13, 2026Contributor·5 tasks·

  ·
  7

• [Jobs] Print a task's logs when log streaming never sees it RUNNING

  #11052·kevinmingtarja opened Oct 13, 2026Collaborator·2 of 3 tasks·

  ·

• [Dashboard] Show the entrypoint and task YAML of cluster jobs

  #11050·zpoint opened Oct 13, 2026Collaborator·10 of 12 tasks·

  ·
  13

• [Core] Compute the Kubernetes pod affinity and bin-packing label in Python

  #11049·kevinmingtarja opened Oct 12, 2026Collaborator·2 of 3 tasks·

  ·
  1

• [Serve] Preserve installed TLS credentials across updates

  #11041·amitgelber opened Oct 12, 2026Contributor·3 tasks·

  ·

• Fix AWS VPC failover across regions

  #11040·amitgelber opened Oct 12, 2026Contributor·3 tasks·

  ·

• [Jobs] Preserve submitted work through DB pool timeouts

  #11039·amitgelber opened Oct 12, 2026Contributor·7 of 11 tasks·

  ·
  2

• [Jobs] Handle cancellation during job initialization

  #11038·amitgelber opened Oct 12, 2026Contributor·4 of 7 tasks·

  ·
  3

• [Core] Keep truncate_long_string within max_length when the text has short words

  #11037·chrikrah opened Oct 12, 2026·2 of 5 tasks·

  ·

• [Jobs] Finish a job cancelled right after a controller claims it

  #11035·cg505 opened Oct 12, 2026Collaborator·2 of 5 tasks·

  ·

• [Jobs] Cancel a launch request submitted while its job is cancelled

  #11034·cg505 opened Oct 12, 2026Collaborator·2 of 5 tasks·

  ·

• [Core] Do not skip cloud teardown when a cancelled launch left resources behind

  #11033·cg505 opened Oct 12, 2026Collaborator·2 of 5 tasks·

  ·