See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/323347624

Network's server monitoring and analysis using Nagios

Conference Paper · March 2017

DOI: 10.1109/WiSPNET.2017.8300092

CITATIONS READS
3 1,126

2 authors:

Renita Johnson N. Edna Elizabeth

Sri Sivasubramaniya Nadar College of Engineering Sri Sivasubramaniya Nadar College of Engineering
4 PUBLICATIONS   3 CITATIONS    24 PUBLICATIONS   73 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Network's server monitoring and analysis View project

Implementation of a CCMT Compression Technique to Improve Performance Metrics in SECNTP-CKM for MANET View project

All content following this page was uploaded by N. Edna Elizabeth on 20 June 2018.

The user has requested enhancement of the downloaded file.

 This full-text paper was peer-reviewed and accepted to be presented at the IEEE WiSPNET 2017 conference.

Network’s Server Monitoring and Analysis

Using Nagios
J. Renita1 and N. Edna Elizabeth2
Electronics and communication Engineering, SSN College of Engineering, Kalavakkam, Chennai-603110
Email: 1 [email protected] 2 [email protected]

Abstract—Network Monitoring is a procedure used to monitor memory, disk queue length, disk usage, network collisions,
a computer network system and notify the network administrator adapter transmit rate, etc. Server monitoring is used by every
in case of any outages. It is necessary to diagnose and report the IT organization in some shape or form. All Monitoring tools
issues that would lead to failure or irregularity in the network.
The performance and usage of a network can also be monitored. run on the SNMP protocol. Open source market includes many
Network monitoring can be done as server and application options like Nagios, Cacti, Zenoss, Zabbix, Open NMS etc.
monitoring. The process of monitoring server’s system resources The basic need for monitoring depends on the type of business
like CPU Usage, Memory Consumption, I/O, Network, Disk that is being carried out by the organization. There has been
Usage, Bandwidth etc. is Server Monitoring. In this paper, server always a competition between open source and commercial
monitoring is done using Nagios, which is an open source tool.
Several nodes are added to the network and the performance solutions but many companies tend to acquire open source
and status of the network are monitored. The data is collected projects by introducing different version of projects for com-
and real time statistics is provided and the performance of the mercial and open source market. Various parameters based
network is analyzed. The main motive is to alert the network deployment, reporting, notifications; triggers, alerts, resource
administrator by methods like SMS or E-mail in case of any usage etc. are being tested [2]. Nagios is the most popular
failure in the network. This helps in securing the network by
alerting the potential issues in real time. monitoring system and is made up with almost all linux
Index Terms—Nagios, Server monitoring, CPU Usage, Memory distributions. There are several other plugins, add-on scripts
consumption. that can be customized and used along with the tool. Nagios
is a light weight program and provides a perfect monitoring
I. I NTRODUCTION tool that can be helpful to monitor all the active protocols and
The external devices that are connected to the network have network devices connected to the network topology. It is also
to be monitored at regular intervals in real time. Monitoring capable of providing real time comprehensive graphs and trend
the network deals with the collection of data to provide real analysis. Cacti tool is found to be a performance monitoring
time statistics and to analyze the performance of the network. tool based on a LAMP stack (Linux/Apache/MySQL/PHP)
When an outage or failure occurs in the network, the network and has RRD (Round Robin Database). The process done in
administrator has to be informed. The network has to be Cacti includes collection, management and display of graphs
secured by alerting the potential issues before they become of the collected data. Some distributions (i.e., Fedora) also
major problem. Methods like SMS, E-mail and Pager can be supply a version in their repositories. However, an important
used to alert the network administrator regarding the failure architecture plugin feature of cacti has to be patched in. Cacti
in the network. use Round Robin Databases (RRD) and MySQL database
The term Network monitoring is used to describe a system technologies to store the collected information. MySQL and
that is helpful in monitoring the network topology continu- PHP are used to provide a graphical, web based interface
ously and finds if there is any jamming, slowdown of system to the RRD databases. Zenoss tool was developed by Bill,
or component failure and immediately notifies the network Erik Dahl and Mark Hinkle. The tool is accomplished of
manager via E-mail, SMS or any other alarms in case of monitoring all devices, servers, network and application inside
any problems. Network monitoring is found to be of no use data center. The core database and the events are stored in My
unless the right things are tracked. The usual areas that are SQL database. The tool comes with an integrated package
examined include bandwidth usage, Server performance and that contains all combined modules. Zabbix was developed by
Application performance. Server monitoring is an important Alexei Vladishev, and was first released in 2001. The current
part of any data center monitoring architecture, but too often stable version of Zabbix is 1.8.3. It can monitor the basic
it becomes an essential process in successfully building out SMTP, HTTP, ICMP services without installation of agents.
a holistic monitoring platform. Server monitoring consists of Zabbix has three core modules for its functioning i) Daemons
monitoring the operating system and its associated hardware ii) Agents iii) Web interface. As name suggests Open NMS,
metrics for servers that run the application. It’s the view of initially a Network Management System and one of the oldest
the world from the perspective of the server, but never from monitoringsoftware in early 2000’s open source leaders were
inside the running processes. Basic server monitoring metrics only Nagios & Open NMS. Open NMS recognizes servers in
include CPU system time, CPU wait time, used memory, free data center and services are linked to the interfaces.

978-1-5090-4442-9/17/$31.00 2017
c IEEE 1904
 This full-text paper was peer-reviewed and accepted to be presented at the IEEE WiSPNET 2017 conference.

In this paper, an approach is presented to monitor the TABLE I

devices in the network. A tool named Nagios is used to NAGIOS O BJECTS AND F UNCTION .
monitor the server and the applications that run in the network.
Object Purpose
The tool can monitor the status of the device and can perform
status check and notifies when there occurs any problem in Hosts Servers or devices being monitored
Host groups Groups of hosts
the network. The use of external plug-in helps in notifying Services Services being monitored
the network manager about the problem in the network. Service groups Groups of services
Section II includes the related works of different authors Time periods Scheduling of checks and notifications
Commands Checking hosts and services notifying contacts
and their proposals. Section III deals with the details about processing performance data event handling
the tool and the methodology used for server monitoring. contacts individuals to alert contact groups
Section IV shows the results and graphs obtained. Section V of contacts
give the conclusion and further enhancement that can be done
to this work.
benefits by providing details on the system status in stan-
II. R ELATED W ORKS dardized way. The author has also concluded that the system
Network’s Server monitoring helps us in understanding the provides data in a more proactive management style instead of
System’s resource usage which can help in improving the reactive management style of system to be used and provides
capacity planning and provides a better end-user performance. a valuable insight into the system’s status in a standardized
Fung Po Tsoa et al., (2016) has presented a survey on way and reduces staff training [12].
managing the server and network resources. The survey also C.H. Philip Yuen et al., (2012) has done a study of the
includes virtual machine allocation and its management to im- real time monitoring network with multiple monitors for large
prove the utilization and cost efficiency of the physical servers. scale applications. A study based on the highly scalable
The author has used Software defined networking principles monitoring devices for distributed applications was done. The
and has discussed about the challenges and opportunities for problem is formulated to construct overlays which are used in
converged resource management [5]. minimizing monitoring delay. A Simple, efficient and scalable
Sihyung Lee et al., (2014) discussed the open problems in monitoring algorithm SMon was introduced in order to reduce
network monitoring and suggested guidelines for the future the monitoring delay and reduces network diameter in real
network monitoring system. The author has analyzed about time and in a distributed manner [8].
the integration of the present technologies with the entire net- Anshul kaushik, (2010) has covered the choice of moni-
work monitoring operations and has also focused on network toring of various servers using SNMP protocol and providing
monitoring technologies [11]. open source solutions. The author has also compared many
Rafiullah Khan et al., (2013) have suggested an effective open source tools like Zabbix, Zenoss, Nagios and Open NMS
and automotive network monitoring system to monitor the and has concluded that Nagios is a good package and is being
network switches and informs the administrator when switch used by masses [3].
goes down. The presented network monitoring system can Antonis Papadogiannakis et al., (2012) have proposed an
easily identify the network problem and its effect and is found approach for improving the runtime performance of a large
to be efficient providing full control over the network [9]. class of CPU and memory intensive passive monitoring appli-
Adam kucera et al., (2013) has suggested that different types cations. The improvement of packet processing performance
of management are done with increasing size of the Building is done in this work by enhancing the locality of code and
Management systems and has concluded that Nagios can be data access. The author has presented a new approach called
utilized for monitoring servers [1]. locality buffering to improve the runtime performance of a
Ahmed D. Kora et al., (2012) has discussed about the open large class of CPU and memory intensive passive monitoring
and adaptable platform that supports fault and configuration applications [4].
management for next generation network and also offers
additional features for application and management function III. M ETHODOLOGY
to enable easy and low cost management of new technologies The selection of a specific software mainly depends on the
and services [2]. services that are being monitored and the goals for monitoring.
T. Michael Silver, (2010) has discussed about the implemen- Nagios is found to have wide range of users and offers higher
tation of monitoring system using an Open Source software functionality than other open source tools. The software is
package to improve the availability of services and to reduce found to have a good history of active development, a large and
the response time when troubles occur. He author has also active user community and a significant number of included
provided details about Nagios to monitor servers and WAN and user contributed extensions. Because of the flexibility of
and has concluded that installation of software takes more time the software design that uses a plug-in architecture, service
and computer resources [6]. checks for library-specific applications can be implemented.
Thomas Davis et al., (2009) has presented a comparative Table I shows the Nagios objects and their purpose [6].
study on Nagios, Cacti and the method of installation and The network administrator has to ensure whether the server is

1905
 This full-text paper was peer-reviewed and accepted to be presented at the IEEE WiSPNET 2017 conference.

Fig. 1. Nagios monitoring.

functioning properly. If the border router or Internet connec-

tion goes down, Nagios will be unable to deliver email alerts
to administrators. This allows an administrator to immediately
begin investigating the problem and get the primary Nagios
server back online.
Fig. 1 explains the process of Nagios monitoring where the
status of the monitored components are collected by the server
and the notifications are send to the administrator through
intranet. The tool is installed in Ubuntu using VM Ware
workstation in ISRO. The credentials have to be supplied in
order to access Nagios. Using Nagios Mail server, SNMP
server, Nagios server etc. can be easily monitored. Initially,
the configuration wizard has to be uploaded and managed in
core configuration manager. Fig. 2 explains the flow chart
of server monitoring process done in Nagios tool. The IP
address of the server is added. Four servers are added to the
configured system of ISRO. The servers are 192.168.100.28,
192.168.100.88, 192.168.100.50 and 192.168.100.19. The IP
address of the server is configured and the configuration
wizard is made to run.
The address and the URL of the primary Nagios server are
supplied. The authentication credentials used to login to the
Fig. 2. Flow chart for server monitoring.
primary Nagios web interface is supplied. Host name has to
be specified for identification purpose. Hence, a specific host
name is provided to the server. The server metrics like ping,
I/O Wait, Web interface etc. is selected for monitoring. The
configuration is finished and new Servers (hosts) are created
for monitoring. The servers are added in order to check their
current working status at regular intervals. The representation
of server in green indicates the server is Ok, Red indicates the
server is Critical and orange indicates the status of the server
is unknown.
The tool is configured in such a way that the error notifi-
cation is send via e-mail. The SNMP address, IMAP address
and the mail address of the network administrator has to be Fig. 3. Importing Nagios in VM ware and Nagios setup.
specified for mail alert to occur. The e-mail address of the
network administrator is given and alert mail is received when from Nagios the process is repeated up to five times. After
the server is critical. A second notification mail is send to that error report is send to the network administrator of ISRO
the network administrator from Nagios to recover the server saying that there is system failure.
that is critical. The e-mail alert is repeated until the server is
IV. R ESULTS AND D ISCUSSIONS
recovered. Fig. 3 explains the proposed methodology. Initially,
the IP address of the servers is added and the protocols for In this section we discuss about the results obtained when
monitoring the server are chosen. The condition of the server Network’s Server was monitored in ISRO, Sriharikota. The
is pinged to the administrator. In case when there is no ping interpretation to the results obtained is also discussed. Nagios

1906
 This full-text paper was peer-reviewed and accepted to be presented at the IEEE WiSPNET 2017 conference.

Fig. 4. Nagios-home. Fig. 6. Host-down.

Fig. 7. Monitoring graph of 192.168.100.50.

Fig. 5. Host-up.

was initially installed and configured and the following results

were obtained.
Fig. 3 shows the initial procedure in which Nagios is
imported in VM ware workstation using Ubuntu. The tool
was accessed using the server 192.168.50.131. Username and
password are given to login to Nagios home. The initial set
up can be made in the Workstation. The network topology is
made by defining the nodes in the directory.
The host and the host group have to be configured and the Fig. 8. Host health.
contacts are added. This is done in order to alert the contacts
in case if there is any failure in the server. The software is
configured in such a way that the entire network is monitored of sufficient bandwidth availability. The monitoring is done at
for every 10 s. Nagios can take 5 re-attempts when the service an interval of four hours. The monitoring process continues
is unavailable. After that error state is updated in the status and the status of the server is checked at regular intervals.
column. Fig. 4 shows the homepage of Nagios where the Fig. 8 shows the host health of the servers that are added to
configuration wizard is made to run. the network. The current status of the hosts and switches added
After the server (host) address is added, the monitoring is to the server is shown in pie diagram. Fig. 9 shows the hyper
done at regular intervals. The servers that are in OK condition map representing the servers that are OK and CRITICAL in
are represented in green. Fig. 5 shows the host up status. the network. The location of servers and switches located in
Fig. 6 shows the servers that are in critical condition. When the network is plotted with respect to distance and their status
the service is unavailable or if there is any error in the server (Ok or Critical).
it is represented as CRITICAL condition. Fig. 10 shows the mine map representing the servers that
Fig. 7 shows the monitoring graph for 192.168.100.50 are added to the network topology. The status grid of Host
server. Here Round trip average, packet loss, round trip max- is shown in tabulation form. Fig. 11 shows the notifications
imum and round trip minimum were monitored at an interval that are given by the Nagios monitoring tool along with the
of 4 h. reason for the failure at a particular interval of time. In this
When traffic is high, rtmax will be high (peak). The time case network admin is contacted.
duration taken for request and response will be high. Details Fig. 12 shows the status of the server that is being moni-
are stored in a database. Packet loss is zero in this host because tored. This includes load time, CPU status, memory and swap

1907
 This full-text paper was peer-reviewed and accepted to be presented at the IEEE WiSPNET 2017 conference.

Fig. 9. Hyper map.

Fig. 13. E-mail notification.

Nagios monitoring can be easily done to the servers that are

connected to the network topology and notifications are sent
via e-mail to the network administrator in case if there is any
network outage.

V. C ONCLUSION AND F UTURE E NHANCEMENTS

In this paper, a network monitoring system is presented that
informs the network administrator in case of any failure in the
network topology. Network monitoring was done in ISRO,
Fig. 10. Mine map. Sriharikota. Server monitoring was given more importance
and the server metrics like CPU status, Memory usage and
load time etc. were continuously monitored and their corre-
sponding status was checked at regular intervals. The tool was
configured in such a way that the network administrator gets
e-mail when there is any system failure. E-mail Notification
is repeated until system recovery is done by the network ad-
ministrator. Performance results were obtained for the servers
that are being monitored.
The performance graphs for the server are obtained only
with the help of plug-in. Hence for further development of the
project, the logs or database of Nagios has to be checked. The
Fig. 11. Notifications.
database has to be built using MYSQL or any database. The
display has to be designed based on the requirement of ISRO.
Acknowledgements: This work is a part of Network
Monitoring used in Indian Space Research Organization,
(ISRO), Sriharikota.
R EFERENCES
[1] Adam kucera, Petr Glos, and Tomas Pitner, “Fault detection in
building management system networks,” in IFAC Proceedings, 2013,
pp. 416–421.
[2] Ahmed D. Kora and Moussa Moindze Soidridine, “Nagios based
enhanced IT management system,” Interntional Journal of Engineering
Science and Technology, vol. 4, no. 3, pp. 818–822, 2012.
Fig. 12. Server status. [3] Anshul kaushik, “Use of open source technologies for enterprise server
monitoring using SNMP,” International Journal on Computer Science
and Engineering, vol. 02, no. 07, pp. 2246–2252, 2010.
and their usage. Server monitoring uses the status of these [4] Antonis Papadogiannakis, Giorgos Vasiliadis, Demetres Antoniades,
server metrics for monitoring. Michalis Polychronakis, and Evangelos P. Markatos, “Improving the
Fig. 13 shows the e-mail notification received from Nagios performance of passive network monitoring applications with mem-
ory locality enhancements,” Computer Communications, vol. 35,
monitoring system. In case if there is any failure in the system pp. 129–140, 2012.
the network administrator gets the notification e-mail. [5] Fung Po Tsoa, Simon Jouet, and Dimitrios P. Pezaros, “Network and
A second mail is received that intimates the network admin- server resource management strategies for data centre infrastructures:
a survey,” Computer Networks, vol. 106, pp. 209–225, 2016.
istrator to recover the fault in the server. This e-mail alert is [6] T. Michael Silver, Monitoring network and service availability with open
repeated until the system is recovered. The results show that source software, Information Technology and Libraries, pp. 8–15, 2010.

1908
 This full-text paper was peer-reviewed and accepted to be presented at the IEEE WiSPNET 2017 conference.

[7] Nicola Bonelli and Stefano Giordano, “Network traffic processing with [11] Sihyung Lee, Kyriaki Levanti, and Hyong S. Kim, “Network monitoring:
PFQ,” IEEE Journal on Selected Areas in Communications, vol. 34, Present and future,” Computer Networks, vol. 65, pp. 84–98, 2014.
no. 6, pp. 1819–1833, 2016. [12] Thomas Davis and David skinnes, “Software monitoring using Nagios,
[8] C. H. Philip Yuen and S. H. Gary Chan, “Scalable real-time monitoring Cacti and Prism,” in CUG Proceedings, 2005, pp. 1–5.
for distributed applications,” IEEE Transactions on Parallel and [13] Zhenqi Wang, Yue Wang, Guangqiang Shao, and Ziyan Guo, “Research
Distributed Systems, vol. 23, no. 12, pp. 1226–1235, 2012. and design of network servers monitoring system based on SNMP,”
[9] Rafiullah Khan, Sarmad Ullah Khan, Rifaqat Zaheer, and Muhammad in Education Technology and Computer Science, ETCS’09. First
Inayatullah Babar, “An efficient network monitoring and manage- International Workshop, 2009, vol. 3, pp. 857–860.
ment system,” International Journal of Information and Electronics [14] Zhenqi Wang, Yue Wang, Guangqiang Shao, and Ziyan Guo, “Research
Engineering, vol. 3, no. 1, pp. 122–126, 2013. and development of monitoring system for network servers,” in 2008
[10] Roger Burton West, “Open source network monitoring software,” ITP 4th International Conference on Wireless Communications, Networking
Journals, no. 124, pp. 3–6, 2000. and Mobile Computing, 2008, pp. 1–3.

1909

View publication stats