- We have 4 sites (2 US , EU, Asia). We ll be having 240, 80, 80 ,80 servers resp. in all location. What is the feasible server architecture. Distributed monitoring or one management server can monitor all locations.
- Any H/W server sizing suggestions for monitoring these many nodes.
Of course it depends on how many services per server, 5 or 30 is changing the picture
That’s a pretty small monitoring size though. You can easily do it with a single server monitoring everything without impacting your network bandwidth.
Standard server, nothing fancy. For example an HP DL360, 2x146Gb HDD, 2Gb RAM and whatever linux you like (RedHat, Fedora, CentOS, etc)
Thus said, it depends if the HUMANS monitoring these systems are all in the same location. If there are IT people in Asia monitoring their stuff and the nagios server is in the US, the interface will be quite slow, and if there is a network cut they won’t have any visibility anymore.
Personally I didn’t go the distributed way with v2.x I might revise this judgement with 3.x that I am preparing the roll out soon.