LIFETIME, purple statuses & check intervals
list Kii Noda
Hi everyone, We're in the process of moving away from Zabbix and have been considering Xymon/Hobbit for this move. We do, however, have a few things we need to clear up. We've been deployng Hobbig on a Debian Lenny box and have a few Debian (Sarge, Etch, Lenny) and a few Fedora clients. We are, at this point, wondering about the reasons behind having a LIFETIME of 30 minutes for checks. The way we see things, we'd rather have this as 6 minutes considering a 5 minute interval for the checks. Can anyone please explain why this has been setup as such and if/how this can be changed? Been going through /etc/hobbit files, read man pages and exercised our Google-Fu but could not come up with a solution. Also, could someone please suggest a reason for not moving interval checks to lower values, such as 1 minute? Zabbix has been spoiling us with very small check intervals, giving us the feeling of staying on top of things. Thank you very much in advance, -- kN
list Greg Hubbard
I guess it depends on how much you want your pager to go off. If you are very lonely, then shortening the time interval might keep your pager hopping. The default LIFETIME value allows a server to reboot without adding alot of "purple" clutter. You have the source code so perhaps you can find out for yourself how difficult it might be to change the default value. If you collect data more often you will have to look at the RRD code as well because it is set up for 5 minute samples. I believe there are some hints in the documentation about how to change to a one minute sample interval. Keep in mind that the client may need more than one minute to complete its workload. Others who know much more than I may wish to add their own comments... GLH
▸
On Mon, Jan 25, 2010 at 12:25 PM, Kii NODA <user-d20081af5452@xymon.invalid> wrote:
Hi everyone, We're in the process of moving away from Zabbix and have been considering Xymon/Hobbit for this move. We do, however, have a few things we need to clear up. We've been deployng Hobbig on a Debian Lenny box and have a few Debian (Sarge, Etch, Lenny) and a few Fedora clients. We are, at this point, wondering about the reasons behind having a LIFETIME of 30 minutes for checks. The way we see things, we'd rather have this as 6 minutes considering a 5 minute interval for the checks. Can anyone please explain why this has been setup as such and if/how this can be changed? Been going through /etc/hobbit files, read man pages and exercised our Google-Fu but could not come up with a solution. Also, could someone please suggest a reason for not moving interval checks to lower values, such as 1 minute? Zabbix has been spoiling us with very small check intervals, giving us the feeling of staying on top of things. Thank you very much in advance, -- kN
--
Disclaimer: 1) all opinions are my own, 2) I may be completely wrong, 3) my
advice is worth at least as much as what you are paying for it, or your
money cheerfully refunded.
list Rich Smrcina
Clients still report back to the server in 5 minute intervals by default. That is part of the standard client reporting protocol that even the clients that I wrote for System z operating systems adhere to. The 30 minute timer is used to alert someone that a client has stopped reporting it's status to the server because it is no longer running or the client is in some way distressed. The nature of 30 minutes and the purple color is rooted in Hobbit (now Xymon) being derived from the Big Brother BTF network services monitor.
▸
On 01/25/2010 12:25 PM, Kii NODA wrote:Hi everyone, We're in the process of moving away from Zabbix and have been considering Xymon/Hobbit for this move. We do, however, have a few things we need to clear up. We've been deployng Hobbig on a Debian Lenny box and have a few Debian (Sarge, Etch, Lenny) and a few Fedora clients. We are, at this point, wondering about the reasons behind having a LIFETIME of 30 minutes for checks. The way we see things, we'd rather have this as 6 minutes considering a 5 minute interval for the checks. Can anyone please explain why this has been setup as such and if/how this can be changed? Been going through /etc/hobbit files, read man pages and exercised our Google-Fu but could not come up with a solution. Also, could someone please suggest a reason for not moving interval checks to lower values, such as 1 minute? Zabbix has been spoiling us with very small check intervals, giving us the feeling of staying on top of things. Thank you very much in advance, -- kN
--
Rich Smrcina Phone: XXX-XXX-XXXX http://www.linkedin.com/in/richsmrcina Catch the WAVV! http://www.wavv.org WAVV 2010 - Apr 9-13, 2010 Covington, KY