
A while back, I converted my monitoring from Nagios and Cacti to Zabbix. I loved Nagios because Nagios is very flexible and reliable. I switched to Zabbix because Zabbix was totally configured via a web browser and Zabbix provided Nagios and Cacti in one product versus two products - Nagios provides the monitoring and Cacti provides trending and graphing.
I have run into a few problems with Zabbix that have been driving me crazy. First, Zabbix will just stop monitoring stuff from time to time. It would seem that Zabbix would stop monitoring stuff if enough stuff went down and I could not quantify what "enough stuff" was, but I had to write a cron job to restart the Zabbix server on a nightly basis to ensure Zabbix cleared any problems caused by "enough stuff". I also had to make sure my Zabbix servers were monitoring each other in case there was a problem with the Zabbix server restart. What a pain!
I really liked Zabbix. Once you figured out the configuration, it was easy to create templates and when you added a device all you had to do was add the associated template and the device was monitored for all the things you wanted the device to be monitored for. Very cool!
I never got the hang of "parent" relationships in Zabbix as the relationships seemed to be tied to an item which was defined in a template and if you modified the template, then you change that for everything using the template.
As I do my migration back to Nagios, I have both Zabbix and Nagios running side by side at the moment. I do notice that Zabbix is alerting me for more issues than Nagios. I wonder if there is a problem with how aggressively I am having Zabbix monitor things versus Nagios. I am not certain what the setting are as for both solutions, I have chosen to use the defaults for the most part.
Anyway, back to Nagios. I do miss the web configuration that Zabbix has, but configuring Nagios is not that hard if you do not mind command line, which I do not. I know there are tools to allow you to configure Nagios with a web interface, but I have chosen to only use those Nagios tools that are supported within Ubuntu for update simplicity. I have too many servers to maintain and to also add keeping track and maintaining software versions when Ubuntu does a good job of that by itself, so if the package was not in Ubuntu, I did not think it was worth the time.
What do I have installed for Nagios, well so far, just Nagios and NagiosGrapher. I have not done any configuration for NagiosGrapher other than restarting Nagios and NagiosGrapher is adding graphs to Nagios screen.
I have not really started to migrate the Zabbix trending and graphing to Cacti yet, but I will get there eventually. I did start some migration to Cacti for some of the stuff I was monitoring in Zabbix, but I have not completed the migration yet.
Comments
Post new comment