Answer the question
In order to leave comments, you need to log in
Monitoring system: RRDtool VS Zabbix?
What is better to use to consolidate statistics from several dozen servers: RRDtool or Zabbix?
Arguments are very welcome! :)
Answer the question
In order to leave comments, you need to log in
zabbix - monitoring with a bang (more than 1000 hosts, services)
cacti (rrd) - network load graphs (for clarity)
munin (rrd) - OS load graphs
you are comparing a hammer and an excavator. What are the "arguments"? rrdtool is just a specific database. Zabbix is a powerful enterprise-level monitoring system, where charting is just one of many functions, and not the most important one.
Zabbix is complex and somewhat redundant in the default configuration. However, it has agents, it is a complete monitoring system. As a frontend to rrdtool, I recommend cacti. Has the beginnings of monitoring, but takes data via SNMP. Well, or plugins to write, which since the days of nagios causes allergies;) If you monitor network devices, it will go and somehow, in essence
zabbix is a monitoring and analysis of hundreds of hosts and services, wide opportunities and additional features.
cacti - convenient SNMP monitoring of dozens and hundreds of pieces of iron such as routers and switches.
munin - detailed and diverse monitoring of dozens of disparate PCs and servers.
Yes, almost everything can be added to Cactus, but, in my opinion, it is most in place where you need to monitor the status of a large data transfer network (well, or a large number of some other more or less the same type of SNMP sensors). If you need to keep an eye on a fleet of computers with different services, with different operating systems, it’s better to install either Menin (up to a hundred hosts, plus or minus pliers) or Zabbix (it’s more difficult, but on more hosts it will be rewarded a hundredfold)
Definitely RRDtool. My statistics are drawn from a dozen and a half servers and a dozen routers. Works fast (running scripts in the crone). Graphs are drawn by cgi-scripts on the fly when accessed (actually, the server is not loaded with unnecessary work, for example, as in mrtg).
I advise zabbix, you will get both statistics and graphs and monitoring with alerts.
rrdtool is just a utility for working with a ring database, nothing more, in its pure form it is of little use, you need a frontend (munin or cacti)
munin is ideal for monitoring the localhost out of the box, installed and immediately you have monitoring of all the main local resources, as soon as the need for customization arises, you need to write plugins, in general, it’s not difficult, but confusing, the main disadvantage is scalability (with dozens of hosts, the traffic grows slightly) and the inability to change the data collection period from the nodes (once every five minutes - I think monitoring is nothing )
cacti - more advanced than munin, you can change the check period, it scales out of the box, but there are too many incomprehensible and non-obvious things in the settings (if something fell off, you don’t immediately catch up with where it fell off), the regular poller is no good, you need to screw a non-native, written in C, after a month of use, I had to throw out this
zabbix solution - a thing! works like clockwork, everything is intuitive, scalable, customizable, documentation for 5+, the main disadvantage is that the database eats up space with spoons) so you need good hardware for the server
Didn't find what you were looking for?
Ask your questionAsk a Question
731 491 924 answers to any question