Changes made since version 4.0.3 prior to version 4.0.4:

🚀 Features

• [SCHED-374] Support ephemeral node states in exporter
  • PR: #2701

🐛 Fixes

• SCHED-1456: Fix vmagent remotewrite in umbrella chart
  • PR: #2672

Contributors:
@theyoprst, @Uburro, @itechdima

Changes made since version 4.1.1 prior to version 4.1.2:

🚀 Features

• SCHED-2006: Support GB300 in Active Checks
  • PR: #2654

🐛 Fixes

• SCHED-1456: Fix vmagent remotewrite in umbrella chart
  • PR: #2672

Contributors:
@ChessProfessor, @github-actions[bot], @theyoprst, @Uburro, @ali-sattari

Changes made since version 4.1.0 prior to version 4.1.1:

🐛 Fixes

• SCHED-1946 Do not set unhealthy for DRAIN+ALLOC status
  • PR: #2635

📦 Dependencies

• Patch 3.0.8
  • PR: #2643

Contributors:
@ChessProfessor, @github-actions[bot]

Changes made since version 4.0.2 prior to version 4.0.3:

🐛 Fixes

• SCHED-1879: Fix job_tmpfs_recreate.sh
  • PR: #2601
• SCHED-1764: Bind-mount libdummy not only on login nodes but also on CPU-only workers in GPU clusters
  • PR: #2599
• SCHED-1946 Do not set unhealthy for DRAIN+ALLOC status
  • PR: #2635

📦 Dependencies

• Patch 3.0.8
  • PR: #2643

Contributors:
@rdjjke, @Uburro, @theyoprst, @github-actions[bot], @ChessProfessor

Changes made since version 3.0.7 prior to version 3.0.8:

🐛 Fixes

• SCHED-1946 Do not set unhealthy for DRAIN+ALLOC status
  • PR: #2635

📦 Dependencies

• Patch 3.0.8
  • PR: #2643

Contributors:
@ChessProfessor

Changes made since version 4.0.2 prior to version 4.1.0:

🚀 Features

• SCHED-1779: Basic IMDS sidecar for VMagent
  • PR: #2554
• SCHED-1649 Enable passive checks for GB300 platform
  • PR: #2559
• SCHED-1650 Use kruise-manager v1.8.3 and kruise-helm-hook v0.1.0 with amd and arm images
  • PR: #2549
• SCHED-1787 Upgrade health-checker
  • PR: #2561
• SCHED-1648: Better support for Block topology
  • PR: #2567
• SCHED-1786 nodeSelector, affinity and tolerations for ActiveChecks Helm
  • PR: #2575
• SCHED-1650 Use nebius-o11y-agent 2.0.10 (arm images)
  • PR: #2571
• SCHED-1838: Rack support in dashboards
  • PR: #2594

🐛 Fixes

• SCHED-1814: Upgrade otel-collectors
  • PR: #2584
• SCHED-1879: Fix job_tmpfs_recreate.sh
  • PR: #2601
• SCHED-1764: Bind-mount libdummy not only on login nodes but also on CPU-only workers in GPU clusters
  • PR: #2599

Contributors:
@ali-sattari, @asteny, @dstaroff, @rdjjke, @Uburro, @theyoprst, @github-actions[bot]

Changes made since version 3.0.6 prior to version 3.0.7:

🐛 Fixes

• SCHED-1879: Fix job_tmpfs_recreate.sh
  • PR: #2601
• SCHED-1764: Bind-mount libdummy not only on login nodes but also on CPU-only workers in GPU clusters
  • PR: #2599

Contributors:
@rdjjke, @Uburro, @theyoprst, @github-actions[bot]

Changes made since version 4.0.1 prior to version 4.0.2:

🚀 Features

• SCHED-1525 Upgrade libnccl-dev and libnccl2 to the 2.30.1-1+cuda13.0~custom1 and pin libcublas-dev-12-9
  • PR: #2579

🐛 Fixes

• SCHED-962: Set finite default SuspendTime for Slurm power-save
  • PR: #2586

Contributors:
@asteny, @Uburro, @github-actions[bot]

Changes made since version 3.0.5 prior to version 3.0.6:

🐛 Fixes

• SCHED-962: Set finite default SuspendTime for Slurm power-save
  • PR: #2586

Contributors:
@Uburro

Changes made since version 4.0.0 prior to version 4.0.1:

🐛 Fixes

• respect power down nodes in sconfigcontroller
  • PR: #2556
• SCHED-1788: Fix inconsistent behaviour with automatic powering up
  • PR: #2560

Contributors:
@itechdima, @github-actions[bot], @Uburro