Re: [Nagios-users] checks per host
On 12 July 2010 18:39, Joel Brooks wrote: > hey gang, > > I'm trying to get a sense of what's normal for the number of checks per > host. > > I'm pushing nagios to a number of servers and the list of things I want to > monitor keeps growing. > > For some servers, I've got > 30 checks - some > 50. > > what is "normal" out there? > > is there a practical limit? The golden rule I use is only monitor something if someone actually wants to know. In practice this means I monitor disk space, memory, cpu and whether the virus checker is ok for almost every WIntel server, but everything else just depends. The support teams soon tell me if they don't want monitoring on something! I guess the most checks I have on any server is about 40 - where the server runs stuff for half a dozen similar accounts and each account needs a handful of checks. The most I have on any 'host' is 94 - where a wireless switch has 90-odd access points connected and I need to know if any of them drop off. There are lots of devices which I just ping - a third party does the detailed monitoring for those - I do the ping to make sure the third party is doing their job okay. Whether all this is 'normal' I can't say for sure. -- This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first ___ Nagios-users mailing list Nagios-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/nagios-users ::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. ::: Messages without supporting info will risk being sent to /dev/null
Re: [Nagios-users] checks per host
Thanks guys. I'm not so worried about the monitoring server. It has lots of head room, and/but I will continue to monitor that. I was mainly interested to know about people's experiences using nagios with a lot of checks per host in terms of usability (web interface), configuration, etc. i.e. would it be of any value to focus on checking multiple items per check rather than having an individual check per monitored item as one person suggested. I'm just wondering what people think of using the tools with 100 (or more) checks per host. ..just general opinion / experiences... also wondering about load from checks on the monitored hosts... i.e. it would suck if monitoring actually caused degradation in performance... does having a lot of checks on a host cause problems? i.e. I've got the basics covered - cpu, memory, disk. now i'm adding a lot of application specific checks - queues, performance counters, log file pattern matches, etc. is there a break point (not on the nagios server, but rather on the clients, or in the UI)? thanks for the replies. J > As pointed out, normal depends on your environment. Some sites only > check a handful > of things per host; some check more. The bigger issue is if your > server can handle > the number of checks and in a timely fashion. Look at nagios parameters > Max_service_check_spread, > max_host_check_spread > as well as the various params for individual services such as > check_interval > > If you have a check that takes a long time to complete, you'll probably > want to > run it less frequently. > > If the client were Unix, not windows, you'd have the luxury of being > able to run > the check on the host in question. This can be done either via cron and > have it > generate a status file that nagios then checks or using NSCA. > On Mon, Jul 12, 2010 at 1:39 PM, Joel Brooks wrote: > hey gang, > > I'm trying to get a sense of what's normal for the number of checks per > host. > > I'm pushing nagios to a number of servers and the list of things I want to > monitor keeps growing. > > For some servers, I've got > 30 checks - some > 50. > > what is "normal" out there? > > is there a practical limit? > > i'm using nagios 3 on centos 5 to monitor windows with nsclient++ 3.8. > > cheers, > > J -- This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first___ Nagios-users mailing list Nagios-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/nagios-users ::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. ::: Messages without supporting info will risk being sent to /dev/null
Re: [Nagios-users] checks per host
As pointed out, normal depends on your environment. Some sites only check a handful of things per host; some check more. The bigger issue is if your server can handle the number of checks and in a timely fashion. Look at nagios parameters Max_service_check_spread, max_host_check_spread as well as the various params for individual services such as check_interval If you have a check that takes a long time to complete, you'll probably want to run it less frequently. If the client were Unix, not windows, you'd have the luxury of being able to run the check on the host in question. This can be done either via cron and have it generate a status file that nagios then checks or using NSCA. On Mon, Jul 12, 2010 at 1:39 PM, Joel Brooks wrote: > hey gang, > > I'm trying to get a sense of what's normal for the number of checks per > host. > > I'm pushing nagios to a number of servers and the list of things I want to > monitor keeps growing. > > For some servers, I've got > 30 checks - some > 50. > > what is "normal" out there? > > is there a practical limit? > > i'm using nagios 3 on centos 5 to monitor windows with nsclient++ 3.8. > > cheers, > > J > > -- > This SF.net email is sponsored by Sprint > What will you do first with EVO, the first 4G phone? > Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first > ___ > Nagios-users mailing list > Nagios-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/nagios-users > ::: Please include Nagios version, plugin version (-v) and OS when reporting > any issue. > ::: Messages without supporting info will risk being sent to /dev/null > -- This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first ___ Nagios-users mailing list Nagios-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/nagios-users ::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. ::: Messages without supporting info will risk being sent to /dev/null
Re: [Nagios-users] checks per host
On Mon, Jul 12, 2010 at 1:39 PM, Joel Brooks wrote: > hey gang, > > I'm trying to get a sense of what's normal for the number of checks per > host. > > I'm pushing nagios to a number of servers and the list of things I want to > monitor keeps growing. > > For some servers, I've got > 30 checks - some > 50. > > what is "normal" out there? > > is there a practical limit? The limit is what your HW can handle and what your people can handle seeing per host. Our teams try to make checks that check multiple elements at once to reduce the numbers of checks per host without loss of alarm granularity - for example, our disk checks check all partitions on a host and allow for rich thresholding - user can specify many thresholds to check per service. Same for critical processes - one check to look for all critical processes and report any that have failed. - Max -- This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first ___ Nagios-users mailing list Nagios-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/nagios-users ::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. ::: Messages without supporting info will risk being sent to /dev/null
Re: [Nagios-users] checks per host
> I'm trying to get a sense of what's normal for the number of checks per > host. If you had a data set of information on Nagios installations you may find an average service checks/hosts ratio but I don't really understand the purpose. People are going to want to monitor different things and at different scales the number of things that they can checked per host may be limited because of the sheer number of hosts and the nagios server(s) ability to process them. If your service checks are all being performed in a timely fashion then I wouldn't worry about checking too many things. If you find yourself wishing you had a nagios alert for something, add a service check and see if the nagios servers(s) are still able to process their checks satisfactorily. -- Kyle -- This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first ___ Nagios-users mailing list Nagios-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/nagios-users ::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. ::: Messages without supporting info will risk being sent to /dev/null
