Zabbix Conference 2013 Agenda

The talk will be about the current state of Zabbix, progress we've made and future plans. I will also give an overview of the new feautures of Zabbix 2.2 and share my thoughts about 2.4.

The managing of environmental variables, including temperature, light and humidity, was once confined to associated research and practices. However, with increasing energy costs, plus a growing awareness of the need to manage and reduce humanity's environmental impact, the managing of environmental variables is becoming increasingly common in domestic and in particular business environments for companies keen on reducing costs and wishing to present a green image. Results of the research suggest that it is possible to use open source software and low cost hardware to develop a high performing, cost effective solution that reduces energy consumption and environmental impact. The research also suggested that the created solution was just “the tip of the iceberg” and that you can, within reason, manage anything anywhere. This presents a whole host of exciting future possibilities.

Perobbix was developed as an extremely informative and reliable test for databases based on Perl. Running as a batch check, it is possible to connect to different database types (primarily for Oracle, but successfully used against MySQL and other databases as well), where a configuration file with multiple queries with different controls on time and conditions of execution. With no limitation for the amount of queries to be executed, one host profile can withhold multiple databases. Using MACROS at profile level and one template to each macro. Highly informative, full of timers, extremely helpful for monitoring and debugging databases and the data within. One can make use of different methods, timeout controls and so forth.

In 2008 Brazil hadn't any community about Zabbix, and the software was not known to most people. What changed in 5 years? How a community initiated by one single man (like in Japan) made the difference? Currently this community has more than 700 members, many lectures at local events, articles in magazines and sites specialized in OpenSource products (or nearly) and building additional functions for Zabbix!

A large infrastructure needs flexible and powerful log analysis tools. Zabbix is the platform to handle alerting, but specific tools are used for the log infrastructure. It's a conceptual presentation, but practical examples are shown as well, for both the OS and applications. We hope you can use the concepts in your environment!

With the introduction of Zabbix 2.0 a major new feature for the Zabbix monitoring system was released upon the world: Low Level Discovery (LLD) was born. LLD gives a Zabbix administrator incredible power over what items will be monitored in an automated fashion. The reduction of administrative overhead for the Zabbix administrators can be reduced dramatically when LLD is actively used. During this presentation we will explain how the LLD mechanism in Zabbix works and we'll show you some examples that demonstrate how you can create your own discoveries.

How to keep your Zabbix configuration up to date at your customers in an easy way and how to make a business out of it ? When you have installed Zabbix to monitor your customers' infrastructure it just starts, organizations grow, change, have different problems and Zabbix needs to be adapted. How can you make sure you can do this for your customers in an easy and secure way with the minimum effort and grow a business model from it?

Real life example of an engineer turned pointy-haired-operations-boss (aka Zabbix ambassador-ship) since Zabbix 1.4. Being a Zabbix evangelist over the years I've run into multiple issues to get the software accepted - both in the engineering departments and management. The idea is to look into these issues and more importantly: to come up with some good answers to them. Engineers know Cacti/Nagios, Zabbix seems complex (and can be for new users!). Management wants uptime and looooves business metrics. Make a 'support triangle': business metrics on top, then cluster aggregates then individual machines (don't care about a burning webserver in a cluster with 100 machines, but care about the cluster CPU spiking suddenly). Training engineers and business end-users. A look into the future, how to integrate cmdb, racktables Zabbix etc.

Zabbix performance degradation issues may be observed when the monitoring load becomes too high. In most cases, the bottleneck is the SQL database. Alternative Zabbix architecture is proposed to reach horizontal scalability with rather low cost. NoSQL solution for storing historical data is described.

DBCs migration process to Zabbix. I want to share ideas, hacks and experiences from our Zabbix setup, mostly focused on how we have tried to transform the “raw” standard Zabbix into a useful tool for our firm. How we tier the different Zabbix functions together with the power of screens and maps and how we handle the different user requirements.

Have you ever wondered "why doesn't this item work on this one host (out of hundreds)"? I did, several times, and I've dug through C code and written small C programs to confirm my hypotheses and I've patched and recompiled Zabbix agent to work on this and that host. Every time it was very time consuming because it required a skill I don't really have (low level programming), so I've rewritten Zabbix agent in a more friendly language — Python. So now, I can view the source code of an item on the host in question, call it from Python interpreter, override it and live happily ever after.

It's always good to prevent downtime of our servers. Although, there are situations when everything is looking good on the backend, but users are experiencing some kind of problems - like GUI errors and applications with snail-like reflexes. It would be nice to get informed by Zabbix before being flushed away by user calls. This talk will be about user emulation, monitoring of interaction with our applications, and extracting some useful information out of it!

In Zabbix, selecting the database system may not be the most important issue. However, if you need to deal with a large amount of monitoring data, the database performance could become a critical issue. We measure the performance of the new Zabbix 2.2 along with the latest version of MySQL and PostgreSQL, and present the results. Also we will show the tuning points of each database when used in Zabbix.

The talk will be about how Zabbix is used to monitoring critical environments in Brazil. SERPRO is a governmental organization with responsibility of looking after the critical environments for several Brazilian governmental organizations and the need to assure their perfect operation for all citizens of Brazil. Specific applications can require to work on platforms without native support of agents or other standard protocols, so we need to use a monitoring tool with support for integrating information from many sources.

A complex infrastructure often calls for a simple, 'at a glance' overview of the system status. We are well aware of the power and flexibility of Zabbix, but can it play well with others? Of course it can. This presentation will show some of the ways in which the data provided by Zabbix can be shown alongside data obtained from other systems in a way that provides real business value, with minimal development effort (some Perl code and Twitter Bootstrap magic).

Large companies use Asset Management tools, a lot of times proprietary, to keep facts about servers and equipment. This presentation shows a uni-directional way to store your assets in both Zabbix and Asset Management tools.

Eltele uses Zabbix as a monitoring service for its customers. This includes infrastructure, servers and all the usual stuff. But in addition, we're doing security and environmental monitoring with specialized hardware, served with Zabbix frontend access to the customers, with personalized dashboards.

Presentation will show practical possibilities to use Zabbix in small devices and SoHo environment, discuss problems of embedding Zabbix into OpenWrt, show the possibility to monitor weather conditions and instruct how to cache results using apache. Next to this, there will be practical example of how to manage and monitor a couple of embedded devices. The last topic will be a proof of concept of using Zabbix as the central point for multi-criteria network analysis.

During the course of our presentation we will show how we are using software implementations of the Zabbix Agent & Zabbix Sender protocols to implement extra functionality or to compliment existing features in our ZBX environment. (Near) realtime graphing of agent items, building an API for application profiling/monitoring, ad-hoc querying of information, compliance scans and custom handling of service restarts in Windows are a few examples of what is possible and what we have currently implemented.

The next version of Zabbix will be released with great new functionality like virtual machine monitoring, improved web scenarios and custom modules. In this talk Pavels will walk you through the most important features developed in Zabbix 2.2.

How Zabbix can be used as engine for an Access Control service, collecting various kind of logs, through its embedded items and external scripts and using a custom developed tool to correlate, show and search the flow of accesses.

"Why on earth was I not notified?!" — ever heard that question from a fellow worker? Setting up notifications can be a challenge; debugging them is cumbersome and requires a good understanding of how Zabbix works. Or were you ever asked for a list of people who would be notified on some event? That question is hard to answer, until the event really takes place. The Action Simulator retrofits the frontend with the capabilities necessary to solve such challenges, while keeping changes to a minimum.