All pull requests

Pull requests

• fix(multipod): fix Pathways teardown false failures and pre-training timeout/retry flakes

  #361·KevinYenky opened Oct 2, 2026·4 tasks done·

  ·
  24

• feat(multipod): implement semantic inference output validation for E2E DAGs

  #338·chiajunglien opened Sep 3, 2026·4 tasks done·

  ·
  1

• Refactor pod log verification to real-time kubectl streaming with timestamp tracking

  #308·RUEI4341 opened Aug 6, 2026·4 tasks done·

  ·
  56

• refator: refactor test group timeout design

  InformationalP2
  #290·lechipai0209 opened Jul 6, 2026·4 tasks done·

  ·
  91

• feat: Test TaskGroupWithTimeout

  #283·Jim-0518 opened Jun 23, 2026·4 tasks done·

  ·

• Fix: Fixed Jetstream_benchmarking_servig DAG

  #276·camiloCienet opened Jun 12, 2026·4 tasks done·

  ·
  40

• Feat: add jobset_healthiness_validation DAG

  #256·chiajunglien opened May 6, 2026·4 tasks done·

  ·
  12

• feat(tpu_observability): Add GKE cluster version manager to automate master and node pool upgrades

  #248·yuna-tzeng opened Apr 28, 2026·4 tasks done·

  ·
  6

• feat: automate image builds with different libtpu version daily and enhance JobSet DAG validations

  P0
  #246·yuna-tzeng opened Apr 23, 2026·3/3·4 tasks done·

  ·
  10

• New dag for pathways Elastic training workload

  WIP
  #215·RUEI4341 opened Feb 12, 2026·4 tasks done·

  ·

• Add TPU Capacity Validation DAG

  P2WIP
  #191·chengpinglin opened Jan 27, 2026·4 tasks done·

  ·
  1

• Add tpu-info-streaming-verification DAG to validate streaming mode refresh frequency

  P2WIP
  #179·chengpinglin opened Jan 16, 2026·4 tasks done·

  ·

• Feat: add jobset_ttr_pod_oom DAG for v6e recovery validation

  P2
  #178·chiajunglien opened Jan 15, 2026·4 tasks done·

  ·
  10

• Add TPU Observability validation for help, version, and process monitoring

  P1
  #165·chengpinglin opened Jan 7, 2026·4 tasks done·

  ·
  27

• feat: Implement strict serial execution

  LargeTo be discussed
  #142·AidenYu1673 opened Dec 10, 2025·4 tasks done·

  ·