立即与支持人员聊天
与支持团队交流

Foglight for Infrastructure 5.9.2 - User Guide

Using Foglight for Infrastructure Monitoring log files with Foglight Log Monitor Monitoring IBM PowerVM environments
Before you begin Managing PowerVM HMC agents Monitoring your PowerVM environment
Advanced system configuration and troubleshooting Reference
Advanced System Configuration Foglight for Infrastructure views Foglight Log Monitor views Rules Metrics
Appendix: Building regular expressions in Foglight

Using Foglight for Infrastructure

Foglight for Infrastructure monitors your infrastructure environment and helps you analyze performance issues affecting various hosts. It may be used by itself to monitor physical hosts, or in combination with other agents to monitor virtual hosts. It contains the following major components:

•
Infrastructure Environment dashboard: provides a visual representation of the hosts being monitored in your environment. For more information, see Exploring the Infrastructure Environment dashboard.
•
Foglight for Infrastructure agents: collect information about the monitored hosts. Agents communicate with the Foglight Management Server using the Foglight Agent Manager. For more information, see: Using Foglight for Infrastructure agents, About the WindowsAgent, About the UnixAgentPlus, About the UnixAgent, and About the MultiHostProcessMonitorAgent.

For more information about Foglight for Infrastructure, see the following topics:

•
Host availability alerting
•
Virtual hosts and reported metrics

Host availability alerting

Foglight for Infrastructure 5.8.5.3 has changed the way it alerts users about host availability. It is now able to identify and report the following three states for a monitored host:

•
MONITORED. This is the default state. It is set on the Host object for each successful collection topology submission.
•
UNMONITORED. It is set on the Host object when the agent fails to authenticate, or is otherwise unable to connect to the monitored host.
•
UNAVAILABLE. It is set on the Host object when agent is configured to validate host availability with ping (by setting the Use ping to validate host availability property for that agent to “True”), and a ping command is unsuccessful. For details about configuring agent properties, see About the WindowsAgent, About the UnixAgentPlus, About the UnixAgent, and About the MultiHostProcessMonitorAgent.
 
* 
NOTE: When the Use ping to validate host availability property is enabled on a UNIX® platform, the sudoer file needs to configured to allow the ICMP process to run with NOPASSWD.

In order to allow for accurate host reporting, the agent is able to submit an unmonitored collection only if it had previously connected to the host during its current activation cycle. When the agent connects to the monitored host, a cached host instance is retained (and refreshed each successful collection cycle) as a reference, should the host become unavailable at a future date.

Virtual hosts and reported metrics

Several Foglight components, including Foglight for Virtualization components (Foglight for VMware, Foglight for Hyper-V, Foglight for OpenStack) and Foglight for Infrastructure report metrics for CPU, Memory, Network, and Disk on a monitored host. However, where both virtualization monitoring agents and Infrastructure HostAgents are set up to collect metrics, those reported metrics would not agree. Virtualization monitoring agent metrics are the more accurate values for monitoring purposes as they come from the virtual hypervisor, for example, an ESX host via vCenter®. vCenter accounts for virtualization and its reported metrics should be considered correct.

There are a number of reasons why the reported infrastructure metrics for a virtual host are not as accurate as those of the virtualization cartridge. A guest OS can only identify the virtual hardware resources allocated to the virtual machine on which it is installed and not the actual total physical hardware resources available to the system. Also, the hypervisor adds overhead which manifests in the reported metrics as increased utilization. For more information, see CPU usage of an application differs in virtual machines compared to physical machines (2032).

If, in addition to the virtualization monitoring agent, you set up an Infrastructure agent to collect core metrics, you run the risk of having two agents reporting different values to the same host topology object. This can create frequent topology changes which impacts Management Server performance.

It is recommended that you disable core IC metric collection (CPU, Memory, Network, Disk) for a guest OS system in favour of the virtualization monitoring agent. You can still use the IC agent to collect process and Windows Service data, if desired.

The following virtualization cartridges are recommended for monitoring virtual hosts:

•
Foglight for VMware
•
Foglight for OpenStack
•
Foglight for Hyper-V

Exploring the Infrastructure Environment dashboard

The Infrastructure Environment dashboard helps you monitor, analyze, and investigate the performance of your infrastructure environment.

It allows you to determine potential bottlenecks in your system’s performance.

For more details, see these topics:

•
Accessing the Infrastructure dashboard
•
Selecting a service
•
Running a report for the Infrastructure Environment
•
Exploring the Monitoring tab
•
Exploring monitored hosts
•
Exploring the FAQ Question Viewer
•
Exploring the Administration tab
自助服务工具
知识库
通知和警报
产品支持
下载软件
技术说明文件
用户论坛
视频教程
RSS订阅源
联系我们
获得许可 帮助
技术支持
查看全部
相关文档

The document was helpful.

选择评级

I easily found the information I needed.

选择评级