Linux Server Monitoring
RedHat Linux Monitoring
Performance Monitoring of Linux Servers is Important
Linux servers are widely used by many enterprises to power critical applications in their data centers. Applications ranging from infrastructure services to databases and web servers to critical applications like SAP and CRM systems are hosted on Linux servers. Performance monitoring of the Linux servers and the applications they host is critical for ensuring that the business operates effectively. Slowdowns can result in frustrated users, loss of productivity and increased support costs for the enterprise.
The Need to Monitor Every Aspect of Linux Server Performance
When a Linux server slows down, administrators have to determine what is causing the slowdown: could it be because of a hardware problem? or could it be because of a malfunctioning application running on the server? or could it be because the server is not sized correctly to handle the workload that it is seeing? If there is a bottleneck on the server, which resource is the bottleneck – is it CPU, or memory, or disk, or network?
To be able to effectively monitor Linux servers, administrators need access to metrics from every layer of the server. They should be able to compare performance across these layers to diagnose exactly where the performance bottleneck lies. While server monitoring is important, administrators also need to be able to look in-depth into application performance. For example, a Java application running on Linux may be taking up all of the CPU resources of the server and slowing down the performance seen by the other applications running on the server. To diagnose and fix this problem, it is important to know why the Java application is taking up CPU – i.e., which thread, which method and which line of code is responsible for the CPU usage.
There are a number of monitoring tools that can provide metrics about the performance of the Linux operating system. When a monitoring tool provides a lot of performance data, but with little intelligence around it, the burden is on the administrator to analyze the metrics and determine what needs to be done to fix the performance problem. Analyzing and interpreting thousands of performance metrics can be time consuming. It also requires a great deal of expertise. What Linux administrators need is a monitoring tool that can assess every aspect of Linux performance, correlate between metrics from different layers, identify the exact cause of a problem and present the results in an easy to use manner, so that even helpdesk staff can use the monitoring tool (and Linux experts are not required for solving routine problems).
Comprehensive Monitoring of Linux Server and Application Performance with eG Enterprise
eG Enterprise is a 100% web-based performance monitoring solution for Linux infrastructures. It includes out of the box support for all the different Linux variants including Red Hat Linux, SuSE Linux, Ubuntu Linux, and CentOS. Using a web-based console, administrators can track the status of their heterogeneous Linux server farms, receive alerts when problems happen, view reports on historical performance, and plan the capacity of their server farms.
To monitor a Linux server, the eG agent software
can be deployed on the server. The agent deployment takes at most a couple of minutes, and as soon as the agent is started, it can start monitoring the Linux server hardware, operating system and application processes with little configuration. Agentless monitoring is also supported. Administrators can pick and choose the servers that have to be monitored with agents (e.g., critical production servers) and those that can be monitored in an agentless manner (e.g., staging servers).
To monitor server hardware, eG Enterprise relies on its integration with native hardware monitoring tools from the hardware vendors – IBM, HP, Dell, etc. Monitoring of the server operating system is based on native commands and scripts. Every aspect of operating system performace – including the usage of key CPU, memory, disk and network resources are tracked. Key operating system logs are also monitored, so administrators can be informed about any alerts or errors in the logs.
The same agent that monitors the Linux operating system can also be used to monitor applications running on the server. Popular public domain applications like Apache web servers, Java application servers like JBoss, Tomcat, MySQL databases, etc., can be monitored. Commercial applications hosted on Linux such as Oracle database servers, WebLogic or WebSphere application servers, enterprise applications like SAP and others can also be monitored using eG Enterprise. For a complete list of platforms supported by eG Enterprise, click here >>>.
Integrated Monitoring of Unix and Windows Servers with eG Enterprise
eG Enterprise uses a unique layer model representation to depict the performance of different layers of a Linux server (e.g., hardware, operating system, network, TCP/IP, application processes, etc.). A similar layer model representation is also used for other operating systems that eG Enterprise supports (e.g., Microsoft Windows 2012, 2008, 2003, 2000, Solaris, AIX, HPUX, OS/400, and OpenVMS). This way, eG Enterprise provides a single pane of glass from where administrators can monitor their heterogeneous multi-vendor data center servers from a single console and with a single unified representation model. This greatly simplifies the monitoring and management of heterogeneous server farms.
The monitoring system is licensed per server OS, and not based on the number of CPU cores or sockets, or based on the applications running on it.
Alerting and Reporting on Linux Performance
Baselines for all the key metrics are pre-defined in eG Enterprise based on industry standard best practices. eG Enterprise can also determine automatic thresholds for other metrics based on an analysis of past history. The actual metrics are compared with the thresholds to determine where the problems lie in the infrastructure. With eG Enterprise in place, Linux administrators can start receiving alerts when a key process fails, a critical event is logged in the server log, or when a disk fills up.
eG Enterprise also provides extensive web-based reporting of all the collected metrics. Without even logging into the web console, administrators can get periodic reports of the performance of their server farm.