Shulou(Shulou.com)06/01 Report--

First, why do you need a monitoring system

In an IT environment, there will be a variety of devices, such as hardware devices, software devices, the composition of the system is also very complex.

A variety of applications constitute a complex IT business system, and it is the responsibility of a company's IT department to ensure the normal operation of these resources. For these applications to run stably, professional IT personnel are needed for design, architecture, maintenance and tuning. In this process, in order to control the availability of the basic environment and business application system in time, it is necessary to obtain the running status of each component, such as CPU utilization, system load, service operation, port connectivity, bandwidth traffic, website access status code and so on. And all of this is inseparable from the surveillance system.

Second, the realization of the monitoring system

The composition of a monitoring system can be divided into two parts: data acquisition part (client side) and data storage analysis alarm display part (server side). These two parts constitute the basic model of the monitoring system.

The working mode of data acquisition can be divided into passive mode (server to client to collect data) and active mode (client actively reports data to server). In general, most monitoring systems should be able to support both modes. The passive mode has a large overhead on the server and is suitable for small-scale monitoring environment, while the active mode has less overhead on the server and is suitable for large-scale monitoring environment.

The protocol methods of data collection can be divided into two types: private client acquisition and public protocol collection (SNMP/SSH/Telnet, etc.).

For the collected monitoring data, it can be stored in the database or text or other means, the specific use of which, should be decided according to the actual needs.

For the general monitoring environment, there are not many nodes to be monitored and less data is generated, so it is enough to use C _ Client/Server (client / server) architecture, which is suitable for small-scale environments in the same region.

For large-scale monitoring environment, there are many nodes to be monitored, and there are many types of monitoring, so the data and network connection overhead generated by monitoring will be very huge, and due to cross-regional and other factors, distributed solutions are needed. The common way is C/P/S (Client/Proxy/Server, client / agent / server) architecture. The use of intermediate agents will greatly improve the processing speed of the monitoring server. Thus it can support the construction of a large-scale distributed monitoring environment.

The more important function of the monitoring system is alarm and fault handling, which is very important for timely problem solving and fault self-healing. When warning, it is necessary to take into account the effective and centralized reporting of faults to prevent "alarm flood", that is, the same kind of alarm information is sent repeatedly and in large quantities.

Third, alarm requirements of the monitoring system

Support a variety of ways, such as SMS, email, IM and other interfaces. It has customizable function and provides programmable interface to third-party alarm media. This is very important in many situations, for example, sending the alarm result to a dedicated alarm analysis system.

Support automatic processing of alarm content analysis to prevent false positives, false positives, and jitter. This is a subject worthy of challenge and research for most monitoring systems. For example, if a computer room network fails, according to the regular warning content, you will receive numerous alarm messages about the failure of each device, while for more advanced alarm messages, we want to receive "there is a network failure in a computer room, the IP of the affected equipment is x.x.x.x~x.x.x.x, and the affected business is xxx." the purpose of doing this is to make the alarm information smarter and more effective. Prevent the occurrence of "alarm".

IV. Current situation of open source software in monitoring system

In monitoring software, open source solutions include traffic monitoring (MRTG, Cacti, SmokingPing, Graphite, etc.) and performance warnings (Nagios, Zabbix, Zenoss Core, Ganglia, OpenTSDB, etc.). Each kind of software has its own characteristics and functions, their respective priorities and goals are not exactly the same, and they are more or less the same in design concepts and implementation methods, but they all share common characteristics, such as Data collection, analysis and display, alarm and simple automatic fault handling. In the end, it can achieve a complete demonstration of the availability of IT system services.

1. Cacti

Cacti is a set of graphical analysis tools for network traffic monitoring based on PHP, MySQL, SNMP and RRDtool. It uses snmpget to get data and uses RRDtool to draw, but users do not need to know the complex parameters of RRDtool. He provides very powerful data and user management functions, can specify that each user can view the tree structure, host device and any picture, can also be combined with LDAP for user authentication, but also can customize templates, in the display and monitoring of historical data, its function is quite good.

By adding templates, Cacti makes the monitoring and addition of impassable devices reusable, has the function of customizable drawing, and has powerful computing power (superposition of data).

2. Nagios

Nagios is an enterprise-level monitoring system, which can monitor the running status and network information of the service, monitor the specified local or remote host parameters and services, and provide abnormal alarm notification function.

Nagios can run on Linux and Unix platforms and provides an optional browser-based Web interface to make it easy for system administrators to view network status, system problems, logs, and so on.

The function of Nagios focuses on monitoring the availability of the service and can alarm in time according to the penalty conditions.

At present, Nagios also occupies a certain market share, but Nagios does not keep pace with the times and can no longer meet the monitoring needs of many times. The expansibility and convenience of the architecture need to be enhanced, and its advanced functions are integrated in the commercial version.

3. Zabbix

Zabbix is a distributed monitoring system, which supports a variety of acquisition methods and clients, has a dedicated Agent, and can also support many protocols, such as SNMP, IPMI, JMX, Telnet, SSH and so on. It says that the collected data is stored in the database, and then analyzed and sorted out to achieve the conditional trigger warning. Its flexible expansibility and rich functions can not be compared with other monitoring systems. Relatively speaking, its overall function is very excellent.

With the large-scale application of cloud computing and virtualization, as well as the rise of mobile Internet and Internet of things in the future, Zabbix will be used more and more widely, and there will be more and more references. At present, many Internet companies, cloud computing companies, system integration software companies, outsourcing service companies and so on, have secondary development and large-scale use of Zabbix.

Zabbix is suitable for users of small and medium-sized enterprises and large and medium-sized enterprises. A single Server node can support tens of thousands of devices, can handle 15000 requests per second, and theoretically can support 50, 000 devices.

Zabbix is an enterprise-level highly integrated open source monitoring software that provides distributed monitoring solutions that can be used to monitor the availability and performance of devices, services, etc. Some of its products are enterprise and community editions, which are truly open source products that users can download and use freely.

Zabbix SIA is the company founded by the official technical team of Zabbix, which operates on the basis of open source commercial software-free of charge for the use of the software and charge for the service. It provides users with consulting and technical support services (customized development, solutions, personnel training, etc.).

Reasons for choosing Zabbix

Compared with similar monitoring products, there are the following reasons to choose to use Zabbix.

1. Zabbix is a free and open source product, users can modify and redevelop the source code at will.

two。 Installation and configuration are simple, and users only need some simple learning to complete the monitoring construction work.

3. The construction environment is simple, and the platform is based on open source software, which only needs Linux, Apache/Nginx, MySQL/PostgreSQL/Oracle and PHP, without special operating system support and dedicated hardware.

4. Zabbix-Agent fully supports the monitoring of Linux, Unix, Windows, AIX, BSD and Solaris. Server and Agent are coded in C language, which takes up very little resources of the system, and the performance and speed of data acquisition are very fast.

5. The data collection is stored in the database for a long time to facilitate the secondary analysis of the monitoring data.

6. Very rich expansion ability, it is easy to customize monitoring items and achieve data collection, and can monitor almost all data. For example, you can monitor the number of visits to the website, monitor UPS and weather temperature, and so on. It is no exaggeration to say that in the world of Zabbix, there are often things you can't think of, and there are no things that can't be done.

7. The mode of operation of the open source community, there are various forums, mailing lists, IM timely communication and so on.

Therefore, if you are a system administrator, network administrator and operation staff, and want to build your own monitoring system environment, Zabbix will be the best choice. If you are a developer and want to develop your own monitoring system based on open source software, Zabbix is also a better choice.

VI. The structure of Zabbix

Zabbix supports monitoring in a variety of network modes, and monitoring agents can be deployed and installed in a distributed manner. The entire architecture is shown in the following figure.

VII. The functional characteristics of Zabbix

Zabbix has the functions of common commercial monitoring software, such as host performance monitoring, network equipment performance monitoring, database performance monitoring, FTP and other general protocol monitoring, multiple alarm methods, detailed storm chart drawing, distributed, scalable, API and so on.

Data collection

L supports Agent, SNMP, IPMI, JMX, SSH, Telnet, etc.

L custom detection.

L customize the frequency of data collection.

Flexible trigger

L can define very flexible alarm thresholds and conditions associated with multiple alarms.

Highly customizable warning

Send notifications, which can be customized including alarm level, action upgrade, recipient and media type.

L notifications can use global macro variables and custom variables.

The automatic processing function includes the automatic calling and execution of remote commands.

Real-time drawing function

The monitoring item draws the data on the graph in real time.

Web monitoring capability

L Zabbix can simulate a browser to request a website and check the return value and response time.

Multiple visual displays

L you can customize the display diagram of monitoring to display a variety of monitoring data in one diagram.

L network topology diagram.

Custom Screens and Slide shows can display a variety of graphics in a centralized manner.

L report function.

L Monitoring and display of resource usage.

Storage of historical data

L data is stored in the database.

L the storage cycle of historical data is configurable.

Delete expired historical data on a regular basis.

It is very easy to configure

The configuration is relatively simple, only the following two steps are needed.

Step 1: add a device.

Step 2: monitoring can be completed by applying the template.

Use templates

L templates can be grouped.

L templates are inheritable.

Network discovery

L supports automatic discovery of network devices and servers.

L Agent automatic registration.

Fast access interface

The Web page is based on PHP.

L remote access.

L log audit.

API function

L the application of API function can be easily combined with other systems, including the use of mobile clients.

System permissions

Disconnected users show different monitoring resources.

L authentication of the user.

Program characteristics

L is written in C language, and its performance and memory overhead are very small.

Support for large-scale environment

L remote monitoring can be easily constructed by using Zabbix-Proxy.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.