Cloud DevOps Admin Guide

Zabbix JMX monitoring setup for Linux server

JMX monitoring can be used to monitor JMX counters of a Java application.

JMX monitoring has native support in Zabbix in the form of a Zabbix daemon called “Zabbix Java gateway”, introduced since Zabbix 2.0.

To retrieve the value of a particular JMX counter on a host, Zabbix server queries the Zabbix Java gateway, which in turn uses the JMX management API to query the application of interest remotely.

What is Java and JMX monitoring?

JMX monitoring can be used to measure all JMX counters of a Java application. Java Management Extensions (JMX) is a Java Community Process (JSR-3) specification for managing and monitoring Java applications. Via a so-called Java gateway, the Zabbix server can address the JMX monitoring services, read data from a Java application and save and process it as an item.

A typical use case is the monitoring of the memory consumption of a Java application or of the Java runtime environment in which the application is operated.

Java Gateway and Java Pollers

Zabbix Java gateway

The so-called Zabbix Java Gateway is a special bollard process that can retrieve data via JMX. Unlike the previously mentioned boll processes, this is not an "internal" process within the Zabbix server. The Java Gateway is a stand-alone daemon that provides data to the Zabbix server through a TCP port. The Java Gateway is Java software that requires a JRE.

Enable Java Gateway

To use JMX monitoring, you must install the Zabbix Java Gateway, which is typically not included in the default installation. If you are using Zabbix LLC DEB or RPM packages , install the Java Gateway as follows. The Java Gateway requires no configuration and can be started immediately.

apt-get install zabbix-java-gateway # Debian/Ubuntu

/etc/init.d/zabbix-java-gateway start # Debian/Ubuntu

yum install zabbix-java-gateway # Red Hat/CentOS

service zabbix-java-gateway start # Red Hat/CentOS

systemctl enable zabbix-java-gateway # Red Hat/CentOS

You should now find startup and shutdown scripts in the /opt/zabbix-java-gateway/sbin/zabbix_java folder .

cd /opt/zabbix-java-gateway/sbin/zabbix_java

./startup.sh

Test if the Java Gateway is running.

# ps -ef| grep -i java

Java Gateway as data provider for the Zabbix server

Now that Java gateway is running, you have to tell Zabbix server where to find Zabbix Java gateway. This is done by specifying JavaGateway and JavaGatewayPort parameters in server configuration file.

If the host on which JMX application is running is monitored by Zabbix proxy, then you specify the connection parameters in proxy configuration file instead.
JavaGateway=<Client IP Address> JavaGatewayPort=10052

By default, server does not start any processes related to JMX monitoring. If you wish to use it, however, you have to specify the number of pre-forked instances of Java pollers. You do this in the same way you specify regular pollers and trappers.
StartJavaPollers=5

Do not forget to restart server or proxy, once you are done with configuring them.

Enter the Java Gateway in the file /etc/zabbix/zabbix_server.conf . In the Zabbix server, activate at least one Java bollard, which "requests" the item data from the Java gateway and forwards the server process. After you make the changes, restart the Zabbix server.

NOTE: This is a one time setup in the zabbix server and need NOT be modified again for every new server that we add to the Zabbix WebUI.

Enabling remote JMX monitoring for Java application

A Java application does not need any additional software installed, but it needs to be started with the command-line options specified below to have support for remote JMX monitoring.

As a bare minimum, if you just wish to get started by monitoring a simple Java application on a local host with no security enforced, start it with these options:

export CATALINA_OPTS="$CATALINA_OPTS -Dcom.sun.management.jmxremote -Dcom.sun.management.jmxremote.port=10052 -Dcom.sun.management.jmxremote.ssl=false -Dcom.sun.management.jmxremote.authenticate=false -Djava.rmi.server.hostname=ClientIP -Djava.net.preferIPv4Stack=true"

NOTE: Communication between Java gateway and the monitored JMX application should not be firewalled.

So disable the server level internal firewall in the linux client using the below command.

systemctl disable firewalld

Configuring JMX interfaces and items in Zabbix frontend

With Java gateway running, server knowing where to find it and a Java application started with support for remote JMX monitoring, it is time to configure the interfaces and items in Zabbix GUI.

Before retrieving data from the JMX interface, you must specify for each host on which IP address and on which TCP port the JMX interface of the Java program listens. Navigate to the host configuration and add a JMX interface. If you want to monitor several Java programs via JMX on a host, you must use different TCP ports.

Apply the appropriate JMX template to the server in the zabbix webUI like JMX generic template and you can see that the JMX turns green and is enabled.

Modern way of /etc/motd - FireMotd for linux

While developing, playing or working on Linux systems, a dynamic MotD generator script can quickly give you an overview of all used components of your Linux systems.

FireMotD can show you this information in a sanitized and colorful way while you log in with SSH or console.
Depending of the chosen theme FireMotD will output all information defined in the theme on your server.
We need EPEL repository to be installed and enabled in the server to install the dependency packages.

Install Dependencies

You need to install the required dependencies as shown below.

yum install bc sysstat jq moreutils

After installing the dependencies, clone/download or copy the FireMotd to the root directory as shown.

git clone https://github.com/OutsideIT/FireMotD.git

Change to the FireMotd directory and run the below commands

You need to have make installed on the system, if you want to use the Makefile.

To install to /usr/local/bin/FireMotD

sudo make install

With this you can probably run FireMotD from anywhere in your system. If not, you need to add /usr/local/bin to your $PATH variable. To adjust the installation path, change the var IDIR=/usr/local/bin in the Makefile to the path you want.

To install bash autocompletion support

sudo make bash_completion

With this you can use TAB to autocomplete parameters and options with FireMotD. Does not require the sudo make install above (system install), but requires the bash-completion package to be installed and working. Then you should logout-login or source the bash completion file, eg. $ . /etc/bash_completion.d/FireMotD

If you don't have root access, just install everything on your user's folder and source the file from your user's .profile file

Crontab to get system information

Root privilege is required for this operation. Only /etc/crontab and the files in /etc/cron.d/ have a username field.

The recommended way to generate /var/tmp/FireMotD.json is by creating a separate cron file for firemotd like this:

sudo vim /etc/cron.d/firemotd

# FireMotD system updates check (randomly execute between 0:00:00 and 5:59:59) 0 0

* * * root perl -e 'sleep int(rand(21600))' && /usr/local/bin/FireMotD -S &>/dev/null

But you can also put it in root's crontab (without the user field):

sudo crontab -e

# FireMotD system updates check (randomly execute between 0:00:00 and 5:59:59) 0 0

* * * perl -e 'sleep int(rand(21600))' && /usr/local/bin/FireMotD -S &>/dev/null

Adding FireMotD to run on login

Choosing where to run your script is kind of situational. Some files will only run on remote logins, other local logins, or even both. You should find out what suits best your needs on each case.

To add FireMotD to a single user

Edit the user's ~/.profile file, ~/.bash_profile file, or the ~/.bashrc file
nano ~/.profile

Add the FireMotD call at the end of the file (choose your theme)

/usr/local/bin/FireMotD -t blue

To add FireMotD to all users

You may call FireMotD from a few different locations for running globally.
Eg./etc/bash.bashrc, /etc/profile.

we use /etc/profile, so add the line to this file.

Color outputs :

Blue

Red

Gray

Enable EPEL Repository for RHEL/CentOS 7.x

What is EPEL

EPEL (Extra Packages for Enterprise Linux) is open source and free community based repository project from Fedora team which provides 100% high quality add-on software packages for Linux distribution including RHEL (Red Hat Enterprise Linux) and CentOS.

Epel project is not a part of RHEL/CentOS but it is designed for major Linux distributions by providing lots of open source packages like networking, sys admin, programming, monitoring and so on. Most of the epel packages are maintained by Fedora repo.

Why we use EPEL repository?

Provides lots of open source packages to install via Yum.
Epel repo is 100% open source and free to use.
It does not provide any core duplicate packages and no compatibility issues.
All epel packages are maintained by Fedora repo.

How To Enable EPEL Repository in RHEL/CentOS 7?

First, you need to download the file using Wget and then install it using RPM on your system to enable the EPEL repository. Use below links based on your Linux OS versions. (Make sure you must be root user).

# wget http://dl.fedoraproject.org/pub/epel/epel-release-latest-7.noarch.rpm # rpm -ivh epel-release-latest-7.noarch.rpm

You need to run the following command to verify that the EPEL repository is enabled. Once you ran the command you will see epel repository.

# yum repolist