“ You Do not manage what you do not measure, you do not measure what you do not define, you do not define what you do not understand, and there is no success in what you do not manage! ”
William Edwards Deming
In computing, one of the most important and often overlooked areas is monitoring the environment. It is very common for professionals to spend thousands of hours deploying hardware and software and expending little or no time to monitor these deployments. After job finished, let's go to the next one. And in almost 100% of cases, something that you do not think about is that you need to monitor these systems.
In small environments that do not have dedicated servers this is "acceptable". However, from the moment you have a dedicated server this is already unacceptable. Monitoring is not only about generating graphics or cute tables for display.
Monitoring is an indispensable tool to monitor the life cycle of an equipment, software or service. Through monitoring, you are able to anticipate problems and / or react proactively, many times, well before the user area realizes that a problem is occurring. It also helps you plan equipment growth or renewal strategies.
Coming from the industrial arena, I have the vision the view that monitoring is part of the process of deploying any solution or service, after all, a stopped machine could represent millions of dollars at a loss. Bringing this vision to the IT environment is already a bit more complicated because we have the culture of acceptable failures introduced with the advent of Windows. Where, if a solution is good enough even if it fails, it will be sold and the flaws will be corrected after the product launches. Therefore, monitoring is often neglected in favor of the increasingly tight deadlines of the business areas of the companies.
Fortunately, this picture has changing, but not yet with the speed it should. If in the industry we had to create dedicated networks to communicate with the devices in expensive devices / converters, through protocols most often proprietary. Networks that are very expensive to deploy and generally very weak, in IT we have the advantage of using ETHERNET networks for practically everything. And we can monitor equipment that connects to TCP / IP networks through some open protocols, but not less efficient, and especially the SNMP,
(Simple Network Management Protocol or Protocolo Simples de Gerenciamento de Redes).
This protocol is used to query a kind of information table called OIDs (Object Identifiers) and through of
MIB (Management Information Base or Bases de Gerenciamento de Informação), some network management system, can interpret this information and display it in the form of numbers, texts, graphs, tables, etc. There are currently 3 versions of this protocol called: V1, V2C and V3. Being version 3 focused on systems in which the security of collected information is a non-negotiable requirement.
There are hundreds of monitoring systems, among them MRTG, PRTG, CACTI, Nagios etc, paid and free, Zabbix being one of the best free. Look here a comparison between them
Zabbix is able to work with the SNMP protocol (among others) to query information on devices that make this protocol available. And one of the advantages is that you can either create your own templates or download ready templates from the Zabbix-Share. However, making your own templates is priceless. In addition to learning about how devices communicate, you will come to understand more deeply about the equipment being monitored,
In this case, I had to learn a lot about the operation of the UNIFI devices to be able to monitor them and it was a very rewarding experience. That I now share with you.
Did you like the text?
Leave your comment or suggestion here.
English 
Português do Brasil