Preface

Network administrators are facing an interesting challenge these days. On the one hand, computer networks are not something new anymore. They have been around for quite a while: their physical components and communication protocols are fairly well understood and don't represent a big mystery to an increasing number of professionals. Moreover, network appliances are getting cheaper and easier to set up, to the point that it doesn't take a certified specialist to install and configure a simple network or connect it to other networks. The very concept of networking is so widespread and ingrained in how users and developers think of a computer system that being online in some form is expected and taken for granted. In other words, a computer network is increasingly seen as a commodity.

On the other hand, the very same forces that are calling for simpler, easier, accessible networks are the ones that are actually pushing them to grow more and more complex every day. It's a matter of both quantity and quality. The number of connected devices on a given network is almost always constantly growing and so is the amount of data exchanged: media streams, application data, backups, database queries, and replication tend to saturate bandwidth just as much as they eat up storage space. As for quality, there are dozens of different requirements that factor in a given network setup: from having to manage different physical mediums (fiber, cable, radio, and so on), to the need to provide high performance and availability, both on the connection and on the application level; from the need to increase performance and reliability for geographical links, to providing confidentiality, security, and data integrity at all levels, and the list goes on.

These two contrasting, yet intertwined, tendencies are forcing network administrators to do more (more services, more availability, and more performance) with less (less budget, but also less attention from the management compared to newer, flashier technologies). Now, more than ever, as a network admin, you need to be able to keep an eye on your network in order to keep it in a healthy state, but also to quickly identify and resolve bottlenecks and outages of any kind—or better yet, find ways to anticipate and work around them before they happen. You'll also need to integrate your systems with different tools and environments (both legacy and strategic ones) that will be out of your direct control, such as asset databases, incident management systems, accounting and profiling systems, and so on. Even more importantly, you'll need to be able to show your work and explain your needs in clear, understandable terms to nontechnical people.

Now, if we were to say that Zabbix is the perfect, one-size-fits-all solution to all your network monitoring and management problems, we would clearly be lying. To this day, no such tool exists despite what many vendors want you to believe. Even if they have many features in common, when it comes to monitoring and capacity management, every network has its own quirks, special cases, and peculiar needs, to the point that any tool has to be carefully tuned to the environment or face the risk of becoming useless and neglected very quickly.

What is true is that Zabbix is a monitoring system powerful enough and flexible enough that, with the right amount of work, can be customized to meet your specific needs. And again, those needs are not limited to monitoring and alerting, but also to performance analysis and prediction, SLA reporting, and so on. When using Zabbix to monitor an environment, you can certainly create items that represent vital metrics for the network in order to have a real-time picture of what's happening. However, those same items can also prove very useful to analyze performance bottlenecks and to plan network expansion and evolution. Items, triggers, and actions can work together to let you take an active role in monitoring your network and easily identify and pre-empt critical outages.

In this book, we'll assume that you already know Zabbix as a general-purpose monitoring tool, and that you also used it to a certain extent. Specifically, we won't cover topics such as item, trigger, or action creation and configuration with a basic, step-by-step approach. Here, we want to focus on a few topics that could be of particular interest for network administrators, and we'll try to help them find their own answers to real-world questions such as the following:

• I have a large number of appliances to monitor and have to keep monitoring data available for a long time due to regulatory requirements. How do I install and configure Zabbix so that it is able to manage effectively this large amount of data?
• What are the best metrics to collect in order to both have an effective real-time monitoring solution and leverage historical data to make performance analysis and predictions?
• Many Zabbix guides and tutorials focus on using the Zabbix agent. The agent is certainly powerful and useful, but how do I leverage in an effective and secure way monitoring protocols that are already available on my network, such as SNMP and netflow?
• Load balancers, proxies, and web servers sometimes fall under a gray area between network and application administration. I have a bunch of web servers and proxies to monitor. What kind of metrics are most useful to check?
• I have a complex network with hosts that are deployed and decommissioned on a daily basis. How do I keep my monitoring solution up-to-date without resorting to long, error-prone manual interventions as much as possible?
• Now that I have collected a large amount of monitoring and performance data, how can I analyze it and show the results in a meaningful way? How do I put together the graphs I have available to show how they are related?

In the course of the next few chapters, we'll try to provide some pointers on how to answer those questions. We discuss as many practical examples and real-world applications as we can around the subject of network monitoring, but more than anything, we wanted to show you how it's relatively simple to leverage Zabbix's power and flexibility to your own needs.

The aim of this book is not to provide you with a set of prepackaged recipes and solutions that you can apply uncritically to your own environment. Even though we provided some scripts and code that are tested and working (and hopefully you'll find them useful), the real intention was always to give you a deeper understanding of the way Zabbix works so that you are able to create your own solutions to your own challenges.

We hope we have succeeded in our goal, and that by the end of the book, you'll find yourself a more confident network administrator and a more proficient Zabbix user. Even if this will not be the case, we hope you'll be able to find something useful in the following chapters: we touch upon different aspects of Zabbix and network monitoring and also discuss a couple of less known features that you might find very interesting nonetheless.

So, without further ado, let's get started with the actual content we want to show you.