DocsICP FilingConsole
官方文档
首页

System load diagnosis

更新时间：
复制 MD 格式
产品详情
我的收藏

Analyze the causes of an abnormal one-minute average system load (the load1 metric) and get suggestions to resolve the issues.

Limits

  • Region limits

    This feature is only available in the Chinese mainland and Hong Kong (China).

  • Permission limits

    If you use a Resource Access Management (RAM) user, ensure your Alibaba Cloud account has granted the AliyunECSReadOnlyAccess and AliyunSysomFullAccess system policies to the RAM user.

  • Operating system limits

    Architecture

    Operating system

    x86 architecture

    • Rocky Linux 9.5

    • Rocky Linux 9.1

    • Ubuntu 20.04

    • Alibaba Cloud Linux 3 Container Optimized Edition

    • Rocky Linux 8.8

    • Ubuntu 22.04

    • Alibaba Cloud Linux 3 Pro

    • Alibaba Cloud Linux 2/3

    • CentOS 7.6 or later, or CentOS 8

    • Anolis OS 7/8

    • Ubuntu 24.04

    ARM architecture

    • Alibaba Cloud Linux 3 Pro

    • Alibaba Cloud Linux 3

Scenarios

Common scenarios for running system load diagnostics:

  • An O&M engineer observes that the system load is abnormally high.

  • The system experiences memory-related issues with networking or I/O within a short period.

Procedure

  1. Log on to the Cloud Monitor console, select the target workspace, and in the left-side navigation pane, click All Features.

  2. On the All Features page, click CloudLens for ECS.

  3. At the top of the page, click System Diagnosis (SysOM).

  4. For Diagnosis Mode, select Node Diagnosis. From the Type of diagnosis list, select Scheduling diagnosis. From the Diagnostic Item list, select System load Diagnosis. Select an Instance ID and click Run Diagnosis.

  5. In the Diagnostic Records section, click View Report.

Diagnostic report

  • Basic Information

    The Basic Information section contains key details of the diagnosis, including the instance ID (resource ID), diagnostic item, diagnostic report ID, and start time.

  • Diagnostic conclusion

    If a load anomaly is detected, this section summarizes the results and provides detailed causes.

  • Diagnostic suggestion

    Provides actionable suggestions based on the diagnostic results.

  • Diagnostic details

    If a load anomaly is detected, this section records the details of the anomaly.

Example

The following example shows how to diagnose a load anomaly caused by an incorrect CPU affinity setting.

Problem description

The container monitoring system detects high system load.

Diagnostic analysis

Use the system load diagnosis feature in the console to diagnose the target ECS instance.

  • Conclusion and suggestion

    The one-minute load average is 16.0, with 17 processes in the Running state. The load is more than 1.5 times the number of CPU cores, which indicates a high-load state. The maximum per-core CPU utilization is 100.0%, while the minimum is 1.0%, a difference of 99.0%. The stress processes (PIDs: 41193, 41186, and 41187) have high CPU utilization.

    • Conclusion: The high load is mainly caused by pressure on CPU computing resources.

    • Suggestion: The analysis suggests that incorrect CPU affinity settings may cause uneven CPU load, with tasks accumulating on a few CPUs.

  • Diagnostic details

    You can view more details in the diagnostic details section.

    • Event Overview: View the overall system load and check whether the kernel is under pressure from system calls, hardware interrupts, software interrupts, or I/O.

    • Load Details: View the distribution of processes in R state and D state to identify stress processes in the R state.

Solutions

Add idle CPUs to the affinity scope or cancel the CPU affinity settings to distribute the load across idle CPUs.

Previous：Scheduling jitter diagnosisNext：Crash diagnosis