Updated on 2025-06-06 GMT+08:00

Node Fault Detection (ModelArts Node Agent)

Description

ModelArts Node Agent is a plug-in for monitoring cluster node exceptions, also, a component for connecting to third-party monitoring platforms. It is installed in each Kubernetes resource pool by default. It is a daemon running on each node and can collect node problems from different daemon processes.

Installing a Plug-in

Certain plug-ins are automatically installed when you create a dedicated resource pool. Currently, you cannot upgrade the plug-ins yourself.

Components

Table 1 Plug-in component

Component

Description

Resource Type

maos-node-agent

Monitor cluster node exceptions and interconnect with third-party monitoring platforms.

DeamonSet

Change History

Table 2 Release history

Plug-in Version

New Feature

6.8.0

Added components for detecting ModelArts node faults and monitoring cluster node exceptions.

6.7.0

Added components for detecting ModelArts node faults and monitoring cluster node exceptions.

6.6.0

Added components for detecting ModelArts node faults and monitoring cluster node exceptions.