Monitoring Agent
The Monitoring Agent is an auxiliary agent program independently developed and open sourced by SurferCloud. By installing the Monitoring Agent in cloud hosts or physical cloud hosts, it can allow resources to work better with the cloud platform monitoring system, expand the monitoring depth of resources, and enrich monitoring indicators (such as memory, disk space, processes, etc.).
1. Version Description
| Update Time | Agent Version | Update Content Description | Notes |
|---|---|---|---|
| 2024.01.22 | v1.2.5 | New: Support collection of idle memory, used memory, available memory indicators, new support of memory usage of GPU on GPU cloud host, memory usage rate of GPU card (only for nvidia's GPU) | |
| 2023.11.06 | v1.2.4 | New: Support for GPU usage rate, ECC correction, driver recognizability index collection | |
| 2023.08.15 | v1.2.3 | New: Support for Rocky 8.5 64-bit, Rocky 9.1 64-bit, Ubuntu 22.04 64-bit, Ubuntu 20.04 64-bit, Ubuntu 18.04 64-bit, High Kernel Ubuntu 18.04 64-bit image | |
| 2023.06.20 | v1.2.2 | New: Support for monitoring /dev/nvme on physical cloud hosts in python3 | |
| 2022.12.07 | v1.2.1 | New: uma collection supports multi-disk partition usage rate monitoring | |
| 2022.09.05 | v1.2.0 | Optimized: the logic of uma memory usage rate collection | |
| 2022.08.15 | v1.1.9 | New: Multi-disk monitoring function for physical machines | |
| 2022.07.30 | v1.1.8 | New: uma can be auto-started after installation | |
| 2022.01.25 | v1.1.7 | New: Include memory ECC error count, disk abnormal (ro) count in physical cloud host, support for centos and ubuntu operating systems | The function of memory ECC error count is only supported in python2 |
| 2021.03.08 | v1.1.6 | New: Support for bare metal version 2.0 | |
| 2019.08.12 | v1.1.5 | Fix: The issue that the kernel version higher than 4.18 cannot be used | |
| 2018.01.03 | v1.1.4 | Fix: A bug that could generate zombie processes | |
| 2017.10.31 | v1.1.3 | New: Support physical cloud collection of GPU temperature | |
| 2017.05.15 | v1.1.2 | Optimized: Agent collection method | |
| 2017.03.07 | v1.1.1 | New: Included physical cloud host disk health status check (0 indicates normal, 1 indicates abnormal), support for centos and ubuntu operating systems | |
| 2016.11.01 | v1.1.0 | 1. New: Support for Intranet report & support for Intranet downloading Agent (Windows version update). 2. Optimized: WIndows version no longer needs to configure public and private keys, it is effective immediately after installation | |
| 2016.05.19 | v1.0.5 | 1. Optimized: The Linux version supports Intranet reporting, uma no longer needs to report data to the public network. 2. Automatic configuration is supported, manual configuration is no longer required | |
| 2016.03.25 | v1.0.2 | Optimized: Linux version tcp connection count collection | |
| 2016.01.07 | v1.0.1 | Fixed: A bug in data collection caused by overly long disk names | |
| 2015.10.29 | v1.0.0 | 1. Optimized: The new version is written in a mix of C and Python, without the need for dependencies such as nodejs. 2. Optimized: Simplified configuration can be reused to avoid copying configurations that other hosts cannot use. 3. Fixed: Memory leak issue. 4. Optimized: Support for automatic updates. 5. New: Monitoring agent for windows operating system supported |
Note：
- If you need to use the Linux Intranet data reporting version, please reinstall the agent; umagent now supports physical cloud hosts. If you encounter errors when starting after installation, please resolve the dependencies according to the error message.
- The monitoring agent only supports kernel versions below 4.18. Using kernels higher than 4.18 with UMA may result in some indicators not being able to be obtained.
2. Preparation for Installing the Monitoring Agent
- The installation process requires system administrator users (such as root, administrator, etc.).
- The installation process needs to be performed in the internal network environment of the SurferCloud cloud host.
3. Red Hat/CentOS entire series
Note： Depending on the default python version of the current system, install the py2 or py3 version of the uma installation package. If the installation fails, try installing another version.
3.1.1 Python2 version of uma installation
64-bit operating system:
wget http://umon.api.service.ucloud.cn/static/umatest/uma-1.2.5-1.x86_64.rpm
rpm -ivh uma-1.2.5-1.x86_64.rpm32-bit operating system:
wget http://umon.api.service.ucloud.cn/static/umatest/uma-1.2.5-1.i386.rpm
rpm -ivh uma-1.2.5-1.i386.rpm3.1.2 Python3 version of uma installation
64-bit operating system:
wget http://umon.api.service.ucloud.cn/static/umatest/uma-py3-1.2.5-1.x86_64.rpm
rpm -ivh uma-py3-1.2.5-1.x86_64.rpm3.2 Start
service uma start3.3 Stop
service uma stop3.4.1 Python2 version of uma uninstall
rpm -e uma3.4.2 Python3 version of uma uninstall
rpm -e uma-py34. Ubuntu/Debian entire series
Note： Depending on the default python version of the current system, install the py2 or py3 version of the uma installation package. If the installation fails, try installing another version.
4.1.1 Python2 version of uma installation
64-bit operating system：
wget http://umon.api.service.ucloud.cn/static/umatest/uma_1.2.5-1_amd64.deb
dpkg -i uma_1.2.5-1_amd64.deb32-bit operating system：
wget http://umon.api.service.ucloud.cn/static/umatest/uma_1.2.5-1_i386.deb
dpkg -i uma_1.2.5-1_i386.deb4.1.2 Python3 version of uma installation
64-bit operating system：
wget http://umon.api.service.ucloud.cn/static/umatest/uma-py3_1.2.5-1_amd64.deb
dpkg -i uma-py3_1.2.5-1_amd64.deb4.2 Start
service uma start4.3 Stop
service uma stop4.4.1 Python2 version of uma uninstall
dpkg -P uma4.4.2 Python3 version of uma uninstall
dpkg -P uma-py35. Physical Cloud Disk Status Monitoring Dependency Package
The current agent version has added a metric for the physical cloud host disk health status check. This metric only returns 0 and 1, 0 indicates the disk is healthy, otherwise, it returns 1.
Installation dependency: Dependency: 1. smartmontools 2. MegaCli64 3. dmidecode 4. hpssacli
