NVIDIA/NVSentinel

View on GitHub

NVSentinel detects and remediates GPU faults on Kubernetes nodes

Stars: 385
Forks: 131
Open beginner issues: 0
Indexed issues: 59
Avg merge: 1d 16h
Merged PRs (30d): 131

Dominant language: Go
License: Apache-2.0
Last GitHub push: Sep 19, 2026
Latest indexed: Sep 19, 2026
Contributing guide: Contributing guide
Code of conduct: Code of conduct
Beginner labels: help wanted

Beginner friendly Everything

Open All

59 open issues indexed

Publish the chart's CRDs as a separate nvsentinel-crds chart Open

area/deployment needs-triage

Difficulty 4/5 3-5 days Newbie friendliness 68/100

NVIDIA/NVSentinel#1882 · 1 comment · Opened Sep 18, 2026
[Feature]: dropped Kubernetes write counters cannot distinguish a lost condition clear from a lost condition set Open

area/fault-management priority/P1

NVIDIA/NVSentinel#1881 · 1 assignee · Opened Sep 18, 2026
[Bug]: mongodb-networkpolicy opens the exporter port 9216 but only to the release namespace, so a Prometheus deployed elsewhere can never scrape it Open

area/fault-management priority/P1

NVIDIA/NVSentinel#1878 · 2 comments · 1 assignee · Opened Sep 18, 2026
KOM: three-state / restart-aware condition handling for NPD-sourced policies Open

area/fault-management needs-triage

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

NVIDIA/NVSentinel#1876 · Opened Sep 18, 2026
[Bug]: manual uncordon leaves cordon-by/cordon-reason/cordon-timestamp labels on the node Open

area/health-monitors priority/P1

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

NVIDIA/NVSentinel#1872 · 2 comments · Opened Sep 17, 2026
[Feature]: Pin CI and UAT to GPU Operator 26.7 Open

enhancement priority/P1

NVIDIA/NVSentinel#1870 · 1 assignee · Opened Sep 17, 2026
[Bug]: Janitor GPU reset overwrites operand node labels without checking their prior values Open

area/health-monitors bug priority/P1

NVIDIA/NVSentinel#1869 · 1 assignee · Opened Sep 17, 2026
[Feature]: Add nvidia.com/gpu.deploy.client node selector for GPU-client DaemonSets Open

area/health-monitors enhancement priority/P1

NVIDIA/NVSentinel#1868 · 1 assignee · Opened Sep 17, 2026
[Feature]: Support DRA-only GPU pods in the preflight injection webhook Open

area/preflight enhancement priority/P1

NVIDIA/NVSentinel#1867 · 1 assignee · Opened Sep 17, 2026
[Feature]: Map DRA-allocated GPUs to pods in metadata-collector for partial drain Open

area/core enhancement priority/P1

NVIDIA/NVSentinel#1865 · 1 assignee · Opened Sep 17, 2026
[Feature]: GPUCluster (DRA) mode support in metadata-collector Open

area/core enhancement priority/P1

NVIDIA/NVSentinel#1864 · 1 assignee · Opened Sep 17, 2026
[Feature]: Auto-detect DCGM hostengine address for GPUCluster mode Open

area/health-monitors enhancement priority/P1

NVIDIA/NVSentinel#1863 · 1 assignee · Opened Sep 17, 2026
[Feature]: GPUCluster (DRA) mode support to the labeler's DCGM discovery Open

area/core enhancement priority/P1

NVIDIA/NVSentinel#1861 · 1 assignee · Opened Sep 17, 2026
[Feature]: GPUCluster (DRA) mode support in labeler Open

area/core enhancement priority/P1

NVIDIA/NVSentinel#1860 · 1 assignee · Opened Sep 17, 2026
[Feature]: GPUCluster DRA mode support in Janitor Open

area/health-monitors enhancement priority/P1

NVIDIA/NVSentinel#1859 · 1 assignee · Opened Sep 17, 2026
[Feature]: allow customDrain to apply to a subset of nodes Open

area/fault-management enhancement needs-triage

Difficulty 4/5 3-5 days Newbie friendliness 66/100

NVIDIA/NVSentinel#1857 · 4 comments · Opened Sep 17, 2026
[Bug]: HEA rule evaluation reads every retained event for the node; time bound cannot use an index Open

area/fault-management priority/P1

NVIDIA/NVSentinel#1851 · 14 comments · 1 assignee · Opened Sep 16, 2026
[Feature]: Parallelise labeler's bootstrap label sweep Open

area/core enhancement priority/P2

Difficulty 4/5 3-5 days Newbie friendliness 48/100

NVIDIA/NVSentinel#1846 · Opened Sep 15, 2026
[Feature]: Release GPU health monitor as a host-native systemd package Open

area/health-monitors needs-triage

Difficulty 5/5 Over a week Newbie friendliness 32/100

NVIDIA/NVSentinel#1834 · 8 comments · Opened Sep 14, 2026
[Feature]: bundled Percona versions move with an NVSentinel release, unguarded and unannounced, and can skip a minor or roll the datastore Open

area/deployment priority/P1

NVIDIA/NVSentinel#1797 · 14 comments · 1 assignee · Opened Sep 10, 2026
[Feature]: Bound PolicyEndpoint growth from NVSentinel's namespace-wide NetworkPolicy at fleet scale Open

area/deployment enhancement priority/P1

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

NVIDIA/NVSentinel#1792 · 1 comment · Opened Sep 10, 2026
[Feature]: GPU reset cannot be extended without forking janitor Open

area/api priority/P1

NVIDIA/NVSentinel#1782 · 7 comments · 1 assignee · Opened Sep 9, 2026
[Feature]: Manage postgres schema in a versioned manner Open

area/deployment enhancement priority/P1

NVIDIA/NVSentinel#1772 · 1 assignee · Opened Sep 8, 2026
[Feature]: Support NVSentinel with GPU operator 26.7 Open

enhancement priority/P1

NVIDIA/NVSentinel#1760 · 1 assignee · Opened Sep 7, 2026
[Bug]: a degrading GPU stalls DCGM until gpu-health-monitor is liveness-killed 49 times, while the underlying ECC degradation is never reported Open

area/health-monitors lifecycle/frozen needs-triage

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

NVIDIA/NVSentinel#1748 · 13 comments · Opened Sep 4, 2026
[Bug]: `InfiniBandCharDeviceCheckIsNotHealthy` set on GB300 nodes Open

area/health-monitors bug priority/P1

NVIDIA/NVSentinel#1716 · 1 comment · 1 assignee · Opened Sep 1, 2026
Run scale test end-to-end Open

area/core area/fault-management priority/P1

NVIDIA/NVSentinel#1705 · 7 comments · 1 assignee · Opened Aug 31, 2026
[Feature]: gpuReset serviceManager cannot express NVML clients outside the GPU Operator namespace Open

area/health-monitors enhancement priority/P2

Difficulty 4/5 3-5 days Newbie friendliness 55/100

NVIDIA/NVSentinel#1691 · 7 comments · Opened Aug 28, 2026
[Feature]: Node Drainer configuration should use k8s selectors Open

area/deployment enhancement priority/P1

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

NVIDIA/NVSentinel#1689 · 3 comments · Opened Aug 28, 2026
[Bug]: GPU Reset fails when Google driver manager is used on GKE Open

area/fault-management bug priority/P1

NVIDIA/NVSentinel#1681 · 4 comments · 1 assignee · Opened Aug 27, 2026
[Bug]: nvsentinel could try to clean up old conditions Open

area/fault-management bug priority/P1

Difficulty 5/5 Over a week Newbie friendliness 42/100

NVIDIA/NVSentinel#1673 · 1 comment · Opened Aug 26, 2026
[Bug]: 1.17 InfiniBandCharDeviceCheck marks nodes unhealthy due to missing /dev/infiniband/iss Open

area/health-monitors bug priority/P1

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

NVIDIA/NVSentinel#1663 · 4 comments · Opened Aug 25, 2026
[Feature]: RecommendedAction names assert VM vs bare metal, but node type is never determined at runtime Open

area/health-monitors priority/P2

Difficulty 4/5 3-5 days Newbie friendliness 48/100

NVIDIA/NVSentinel#1661 · 2 comments · Opened Aug 25, 2026
[Feature]: Deploy a MongoDB connection proxy to keep connection count constant as fleet scales Open

area/core enhancement priority/P2

NVIDIA/NVSentinel#1595 · 2 comments · 1 assignee · Opened Aug 14, 2026
[Bug]: MongoDB inherits the host nofile soft limit which may lead to crashes at high node scale Open

area/core bug priority/P1

NVIDIA/NVSentinel#1555 · 1 comment · 1 assignee · Opened Jul 31, 2026
[Feature]: Support recovery events for health-events-analyzer derived conditions Open

area/health-monitors enhancement priority/P1

Difficulty 5/5 Over a week Newbie friendliness 45/100

NVIDIA/NVSentinel#1553 · 6 comments · Opened Jul 31, 2026
[Bug]: Missed delete events will prevent kubernetes-object-monitor recovery Open

bug help wanted priority/P1

Difficulty 4/5 3-5 days Newbie friendliness 48/100

NVIDIA/NVSentinel#1542 · Opened Jul 27, 2026
[Epic]: Scale testing NVSentinel Open

area/core area/deployment area/fault-management area/plugins area/preflight area/tests priority/P1

NVIDIA/NVSentinel#1511 · 2 assignees · Opened Jul 21, 2026
[Feature]: metadata-collector: sysfs-based NIC topology fallback when `nvidia-smi topo -m` fails (e.g. SoftiWARP device present) Open

area/health-monitors enhancement priority/P2

NVIDIA/NVSentinel#1485 · 1 assignee · Opened Jul 13, 2026
[Bug]: Fault-quarantine does not apply cordon or taint actions from additional unhealthy events Open

area/fault-management bug External DSX priority/P1

NVIDIA/NVSentinel#1439 · 1 assignee · Opened Jun 30, 2026
[Feature]: Add optional enrichment pipeline to event-exporter for Pod metadata and business context Open

area/fault-management enhancement External DSX priority/P1

NVIDIA/NVSentinel#1434 · 8 comments · 1 assignee · Opened Jun 29, 2026
[Feature]: Add post-remediation validation gate before clearing node/GPU fault state Open

area/fault-management enhancement External DSX priority/P1

NVIDIA/NVSentinel#1427 · 13 comments · 1 assignee · Opened Jun 26, 2026
[Bug]: NVSentinel doesn't work in IPv6-only clusters Open

area/health-monitors bug External DSX priority/P1

Difficulty 4/5 3-5 days Newbie friendliness 25/100

NVIDIA/NVSentinel#1407 · 6 comments · Opened Jun 18, 2026
[Feature]: Parallelism-aware preflight checks Open

area/preflight enhancement External DSX priority/P2

Difficulty 5/5 Over a week Newbie friendliness 42/100

NVIDIA/NVSentinel#1354 · 8 comments · Opened Jun 1, 2026
[Feature]: Support out-of-tree CSP Health Monitor/Janitor plugin Open

area/health-monitors enhancement External DSX priority/P1

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

NVIDIA/NVSentinel#1336 · 3 comments · Opened May 26, 2026
[Feature]: NIC Health Monitor support for GCP TCP Express Offload (TCPXO) Open

enhancement External DSX help wanted priority/P1

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

NVIDIA/NVSentinel#1326 · 2 comments · Opened May 21, 2026
[Feature]: NIC Health Monitor support for AWS Elastic Fabric Adapter (EFA) Open

enhancement External DSX help wanted priority/P1

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

NVIDIA/NVSentinel#1325 · 2 comments · Opened May 21, 2026
[Feature]: Compact structured representation for active node condition entries Open

enhancement External DSX priority/P1

NVIDIA/NVSentinel#1305 · 1 assignee · Opened May 18, 2026
[Feature/Enhancement]: Coordinate node ownership with external remediation systems Open

enhancement External DSX priority/P1

NVIDIA/NVSentinel#1276 · 1 assignee · Opened May 11, 2026
[Feature/Enhancement]: Fabric health signals for GB*/NVL72 rack-level architectures Open

enhancement External DSX priority/P1

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

NVIDIA/NVSentinel#1275 · 4 comments · Opened May 11, 2026

NVIDIA/NVSentinel

Get new issues in your inbox