What is ack-node-problem-detector? - Container Service for Kubernetes

ack-node-problem-detector detects node anomalies, provides an event center, and supports integration with third-party monitoring platforms for ACK clusters.

Overview

Built on the open-source Node Problem Detector (NPD), this component monitors node health and serves as an event center. It includes:

kube-event-init: Initializes the Simple Log Service (SLS) resources for the event center during installation, enabling ack-node-problem-detector-daemonset and kube-eventer to store, compute, and analyze event data.
ack-node-problem-detector-daemonset: Runs a pod on each eligible node to monitor node health and report cluster conditions and events. In this topic, the ack-node-problem-detector image address refers to the image for ack-node-problem-detector-daemonset.

Note
See the node-problem-detector open-source project.
kube-eventer: Reports cluster events to the SLS event center by default, providing event storage and analysis with a 90-day retention period, dashboards, alerting, and search. You can also configure kube-eventer to forward events to other systems such as DingTalk or EventBridge. See kube-eventer.
accel-health-monitor: Runs a pod on each eligible GPU node to monitor GPU device status and report node conditions and Kubernetes events. Image addresses are listed in the release notes below. For permissions and precautions, see GPU fault detection.

Usage

Event monitoring covers installation, use cases, and plugin features.

Release notes

May 2026

Version	Image address	Release date	Description
1.2.35	`kube-eventer`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer:v1.2.14-4b806cb-aliyun `node-problem-detector`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/ack-node-problem-detector:v0.8.17-952071f-aliyun `accel-health-monitor`: registry-__ACK_REGION_ID__-vpc.ack.aliyuncs.com/acs/accel-health-monitor:v0.5.5-ad7ad729-aliyun	May 18, 2026	Note This version is in a canary release. To use this version, submit a ticket. Added the configuration-triggered mechanism for GPU fault detection and automatic fencing. See GPU fault detection and automatic fencing mechanism. See [Product Change] Announcement on Changes to GPU Fault Detection and Automatic Fencing.

February 2026

Version	Image address	Release date	Description
1.2.30	`kube-eventer`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer:v1.2.14-4b806cb-aliyun `node-problem-detector`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/ack-node-problem-detector:v0.8.17-952071f-aliyun `accel-health-monitor`: registry-__ACK_REGION_ID__-vpc.ack.aliyuncs.com/acs/accel-health-monitor:v0.5.4-4c80dfa0-aliyun	February 2, 2026	Note This version is in a canary release. To use this version, submit a ticket. Applied security hardening to `ack-node-problem-detector-daemonset`. Applied security hardening to `kube-eventer`. Added a console option to enable or disable GPU fault isolation file generation. Changed the fencing strategy for some GPU detection items. See GPU fault detection. Added support for eRDMA detection.

November 2025

Version

Image address

Release date

Description

1.2.29

accel-health-monitor: registry-__ACK_REGION_ID__-vpc.ack.aliyuncs.com/acs/accel-health-monitor:v0.5.3-bafb2ba5-aliyun
kube-eventer: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer:v1.2.14-315a7cb-aliyun

November 30, 2025

Note

This version is in a canary release. To use this version, submit a ticket.

Decoupled the GPU detection plug-in from ack-node-problem-detector-daemonset into a separate DaemonSet, ack-accel-health-monitor. For required permissions, see GPU fault detection.
Added software and device detection to the GPU plug-in, including nvidia-persistenced, nvidia-fabricmanager, and NVLink.
Fixed occasional GPU plug-in restarts caused by JSON serialization failure.
Enabled kube-eventer to report data to SLS over HTTPS.

July 2025

Version	Image address	Release date	Description
1.2.27	`kube-eventer`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer:v1.2.13-b4a3960-aliyun `kube-event-init`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer-init:v1.9-2b115d6-aliyun	July 24, 2025	Note This version is in a canary release. To use this version, submit a ticket. Applied security hardening to `kube-eventer` and `kube-event-init`. ACK Dedicated Clusters now support the hardened mode for ECS instance metadata access, enhancing authentication security. See Enforce the use of hardened mode to access ECS instance metadata.

June 2025

Version	Image address	Release date	Description
1.2.26	`ack-node-problem-detector`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/node-problem-detector:v0.8.16-8d2193b-aliyun `npd-gpu`: registry-__ACK_REGION_ID__-vpc.ack.aliyuncs.com/acs/npd-gpu-plugin:v0.4.1-7359b830-aliyun kube-eventer: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer:v1.2.12-c7c1896-aliyun `kube-event-init`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer-init:v1.8-e43647f-aliyun	June 11, 2025	Note This version is in a canary release. To use this version, submit a ticket. Fixed an issue where the `NvidiaDeviceRecovered` event was not correctly emitted in some GPU self-healing scenarios. Optimized the image size of `ack-node-problem-detector`.

Version	Image address	Release date	Description
1.2.25	`ack-node-problem-detector`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/node-problem-detector:v0.8.16-8ed7053-aliyun `npd-gpu`: registry-__ACK_REGION_ID__-vpc.ack.aliyuncs.com/acs/npd-gpu-plugin:v0.4.0-e434dc36-aliyun `kube-eventer`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer:v1.2.12-c7c1896-aliyun `kube-event-init`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer-init:v1.8-e43647f-aliyun	June 6, 2025	Note This version is in a canary release. To use this version, submit a ticket. Added the `npd-gpu` container for GPU fault detection. Added support for fencing specific GPUs when a fault is detected. Added detection items including `NvidiaXID44Error`, `NvidiaXID61Error`, `NvidiaXID62Error`, and `NvidiaXID69Error`. See GPU fault detection and automatic fencing. Added support for configuring which GPU detection items to enable by using `ack-node-problem-detector-config`. Optimized the image size of `ack-node-problem-detector`.

August 2024

Version	Image address	Release date	Description
1.2.20	`ack-node-problem-detector`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/node-problem-detector:v0.8.14-3c6002c-aliyun `kube-eventer`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer:v1.2.11-0620284-aliyun `kube-event-init`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer-init:v1.8-e43647f-aliyun	August 20, 2024	Added support for GPU fault inspection on ECS nodes. Upgraded `kube-eventer` to resolve performance bottlenecks when reporting large volumes of events. Upgraded `kube-eventer` to support the V4 signature algorithm for SLS data transmission. Added a parameter to configure the local port (20256 or 20257, disabled by default) for the `ack-node-problem-detector` DaemonSet pod.

December 2023

Version	Image address	Release date	Description
v1.2.18	`ack-node-problem-detector`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/node-problem-detector:v0.8.13-003ac31-aliyun `kube-eventer`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer:v1.2.8-27a468a-aliyun `kube-event-init`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer-init:v1.7-48a2acc-aliyun	December 18, 2023	Fixed false positive `podOOMKilling` events caused by cached historical kernel logs. Enabled inheritance of custom component parameters when upgrading from older versions of `ack-node-problem-detector`.

August 2023

Version	Image address	Release date	Description
v1.2.17	`ack-node-problem-detector`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/node-problem-detector:v0.8.12-bf8aff8-aliyun `kube-eventer`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer:v1.2.8-27a468a-aliyun `kube-event-init`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer-init:v1.7-48a2acc-aliyun	August 24, 2023	You can now update SLS Project and Logstore configurations from the Add-ons page in the ACK console. Added support for custom labels, such as a cluster name, when sending log data to SLS. These labels appear by default in the ACK event center.

June 2023

Version

Image address

Release date

Description

v1.2.16

ack-node-problem-detector: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/node-problem-detector:v0.8.12-bf8aff8-aliyun
kube-eventer: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer:v1.2.8-019546c-aliyun
kube-event-init: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer-init:v1.7-48a2acc-aliyun

June 27, 2023

Added support for configuring component resource specifications on the ACK console Add-ons page.

v1.2.15

ack-node-problem-detector: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/node-problem-detector:v0.8.12-bf8aff8-aliyun
kube-eventer: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer:v1.2.8-019546c-aliyun
kube-event-init: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer-init:v1.7-48a2acc-aliyun

June 6, 2023

Reduced API server and etcd load from frequent podOOMKilling events in large clusters.

February 2023

Version	Image address	Release date	Description
v1.2.14	`ack-node-problem-detector`: registry.aliyuncs.com/acs/node-problem-detector:v0.8.11-edc7907-aliyun `kube-eventer`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer:v1.2.6-bbf76f7-aliyun `kube-event-init`: registry.{ .Values.controller.regionId }.aliyuncs.com/acs/kube-eventer-init:v1.7-48a2acc-aliyun	February 3, 2023	Improved the image pull speed for the component. Added support for ACK Edge clusters.

September 2022

Version	Image address	Release date	Description
v1.2.11	`ack-node-problem-detector`: registry.aliyuncs.com/acs/node-problem-detector:v0.8.11-edc7907-aliyun `kube-eventer`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer:v1.2.6-bbf76f7-aliyun `kube-event-init`: registry.{ .Values.controller.regionId }.aliyuncs.com/acs/kube-eventer-init:v1.7-48a2acc-aliyun	September 30, 2022	Optimized `ack-node-problem-detector` inspection logic to reduce load on core cluster components. Applied image security hardening.

February 2022

Version	Image address	Release date	Description
v1.2.9	`ack-node-problem-detector`: registry.aliyuncs.com/acs/node-problem-detector:v0.8.10-e0ff7d2 `kube-eventer`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer-amd64:v1.2.6-f0efecf-aliyun `kube-event-init`: registry.{ .Values.controller.regionId }.aliyuncs.com/acs/kube-eventer-init:v1.6-a92aba6-aliyun	February 22, 2022	Added support for kernel inspection. Applied security hardening.

January 2022

Version	Image address	Release date	Description
v1.2.8	`ack-node-problem-detector`: registry.aliyuncs.com/acs/node-problem-detector:v0.8.10-e0ff7d2 `kube-eventer`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer-amd64:v1.2.5-cc7ec54-aliyun `kube-event-init`: registry.{ .Values.controller.regionId }.aliyuncs.com/acs/kube-eventer-init:v1.6-a92aba6-aliyun	January 20, 2022	Added compatibility for different modes of Containerd. Optimized QoS resource limits to improve component stability.

November 2021

Version	Image address	Release date	Description
v1.2.7	`ack-node-problem-detector`: registry.aliyuncs.com/acs/node-problem-detector:v0.8.10-e0ff7d2 `kube-eventer`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer-amd64:v1.2.5-cc7ec54-aliyun `kube-event-init`: registry.{ .Values.controller.regionId }.aliyuncs.com/acs/kube-eventer-init:v1.6-a92aba6-aliyun	November 25, 2021	Added compatibility with system services on operating systems such as Alibaba Cloud Linux 3 and CentOS 8. Added support for the ARM architecture.

April 2021

Version	Image address	Release date	Description
v1.2.5	`ack-node-problem-detector`: registry.aliyuncs.com/acs/node-problem-detector:v0.6.3-28-160499f `kube-eventer`: registry-vpc.__ACK_REGION_ID__.aliyuncs.com/acs/kube-eventer-amd64:v1.2.4-0f5aaee-aliyun `kube-event-init`: registry.{ .Values.controller.regionId }.aliyuncs.com/acs/kube-eventer-init:1.5-5e0e7c1-aliyun	April 25, 2021	Fixed a "414 Request Too Large" error from `kube-event-init` in `kube-system` when the event center was enabled. Optimized the `kube-eventer` list-watch mechanism to reduce request traffic to etcd. See eventer list-watch. Fixed incorrect timestamp parsing for some system events in `kube-eventer`. See fix FailedScheduling event write to sls with wrong timestamp.

July 2020

Version	Image address	Release date	Description
v0.6.3-28-160499f	registry.aliyuncs.com/acs/node-problem-detector:v0.6.3-28-160499f	July 27, 2020	Enhanced `OOMKilling` event messages to include the pod name, namespace, and UID. Improved the execution efficiency of the `check_fd` plug-in. Improved event notifications for node PID usage thresholds. Upgraded the network problem detection plug-in. Added a plug-in to trigger alerts for node system disk inode usage.