wfsaxton wrote:New (potential) user of Nagios here looking to see whether we should replace our home-grown solution for monitoring 10kish servers/devices.
Welcome.
How does Nagios handle monitoring/configuration large groups of systems? For example, say we have a group of servers we want certain checks on, other servers different checks on, some of these servers using different check "thresholds", etc.
Nagios handles things quite well.
NASA Ames Research Center uses it to monitor, well...you know...NASA stuff like their supercomputer cluster (among other things). At the other end of the scale,
a beekeeper uses it to monitor a beehive. (Shameless plug: You can see both of these people [
plus me] speak at the 2014 Nagios World Conference in October. Check out my signature for a discount link!)
- Is it just based on creating a group of systems and configuring that "group" to run certain checks/use certain thresholds?
- What about hierarchies of these groups?
- What about dynamically creating these groups based on criteria? (Our current home-grown mechanism allows us to configure things based on tags or attributes we have assigned to them; wondering if that is possible with Nagios.)
Your group idea is correct. The specifics may vary. Heirarchies aren't really allowed, but you can put a single machine into as many groups as you would like and only assign checks to those groups that they need to run on.
Dynamically creating things is simplistically complex. Nagios Core runs on a set of configuration files that have a known format. If you can regenerate them on the fly, you can dynamically generate your configs. Chef and Puppet are two examples of how Nagios integrates wonderfully with system management software, but you can probably do the same with Ansible and other types of "thigies."
If you need specific help, my company provides Naigos consulting services beyond the scope of this forum and we'd be happy to help.
