Ah, yeah. Networking performance is rated to be bad on the micro instances. You want at least a medium for anything serious, if only for network performance.
As for Zabbix vs Nagios, I think they're both good at different jobs. Zabbix is great at visualizing and analyzing metrics; if you want to be able to analyze performance trends and such that Nagios just can't dig into (like active users over time doubling within a given time period, but still below a threshold), Zabbix is definitely the way to go. However, in my experience with Zabbix (pre-2.0, however, so caveat that my experience is old and possibly outdated), Nagios is better for alerting and monitoring checks. The Zabbix agent isn't (or wasn't) as robust or as supported as NRPE is. And one issue I ran into with Zabbix is that the equivalent of a Nagios UNKNOWN state actually did not send alerts by default. You actually had to custom trap for that on every check. So, for example, if you had checks that ran over the Zabbix agent, and the agent went offline, the checks would fail but not alert. You actually had to explicitly add a new check to verify that the agent was correctly working. It's also harder to automate configuring something like Cacti or Zabbix that store their configs in a database vs config files like Nagios. If you're doing things manually, that's probably a non-issue. And if you can hook into notifying Zabbix when you spin up a new instance, I believe it supports registration... not sure if that works with something like Chef, but we couldn't do so back when we were using Puppet (although Puppet has also advanced significantly since that time, so that may also be outdated info as well). -- ~*~ StormeRider ~*~ "Every world needs its heroes [...] They inspire us to be better than we are. And they protect from the darkness that's just around the corner." (from Smallville Season 6x1: "Zod") On why I hate the phrase "that's so lame"... http://bit.ly/Ps3uSS On Wed, Oct 8, 2014 at 6:38 PM, Edward Ned Harvey (lopser) < lop...@nedharvey.com> wrote: > > From: tech-boun...@lists.lopsa.org [mailto:tech-boun...@lists.lopsa.org] > > On Behalf Of Lawrence K. Chen, P.Eng. > > > > My first thought, run nagios somewhere...because I have a nagios server > at > > Are you saying you think nagios is better than zabbix? I mentioned this > is the zabbix server (maybe you missed that.) > > After speaking to a whole bunch of people about a whole bunch of systems, > I'm pretty confident that we know what's going on - Simply, the machine in > question is a "tiny" instance, so when Amazon gets enough demand for higher > paying customers on higher paying instances, they take the little machines > like this, and simply flush us down the toilet. It's a big F-U to little > customers. > > Ironically, we're not a little customer. We're a medium customer. Total > monthly usage is several hundred dollars per month, but we have this *one* > tiny instance... > _______________________________________________ > Tech mailing list > Tech@lists.lopsa.org > https://lists.lopsa.org/cgi-bin/mailman/listinfo/tech > This list provided by the League of Professional System Administrators > http://lopsa.org/ >
_______________________________________________ Tech mailing list Tech@lists.lopsa.org https://lists.lopsa.org/cgi-bin/mailman/listinfo/tech This list provided by the League of Professional System Administrators http://lopsa.org/
