J.C.
I looked at tasks.cfg and see the interval for xymonnet. That is 5 minutes. But right below that setting there is one for xymonentagain. This is set to 1 minute for failed systems.
We are trying to test if we can connect to the router so if the ping fails I can only assume that xymon will run the ping test a second time one minute later and then every minute until the unit recovers. If that is correct than it is fine. I did not see if I could only set the ping test to run every minute just for that one unit.
Would the critical systems view be more appropriate for this?
Thanks for any insight into this.
On Sun, Apr 10, 2016 at 9:15 PM, J.C. Cleaver <cleaver at terabithia.org> wrote:
On Sat, April 9, 2016 8:02 am, john boris wrote:
I have searched for changing the interval time between tests. I have a system I need to monitor and they want to have it checked for connectivity every minute. I only need it for that host. I can't find anywhere in the documentation which states how oftent the tests are run or how you change the interval.
-- John J. Boris, Sr.
John,
In a simple cases, default install and no custom tests, adjust the "INTERVAL" value for [xymonnet] in tasks.cfg on your xymond server, and then the [xymonclient] sections in clientlaunch.cfg on your client systems... It's actually not as simple a question as it might seem at first, though :)
Xymon operates with a push-based model, so there's no "central scheduler" with which to schedule test intervals. They're a function of how often the various testing sub-systems run, and/or generate reports. The client status pages (disk space, memory, cpu, etc.) are generated immediately whenever client *data* is received, which is pushed up at that interval from the clients themselves. Network polling occurs when the xymonnet program is launched and reports back.
In simple systems, that's basically it. But Xymon can be easily customized by admins at larger sites to handle all sorts of cases like inverse message retrieval, multiple xymonnet servers, multiple copies of xymonnet at different intervals, custom tests from any source running at different intervals, etc.
Think of testing intervals as more an emergent property of the system as a whole; what you're doing is setting up a message bus upon which test results are received from arbitrary generators at desired intervals.
HTH, -jc Xymon, unlike many other monitoring systems, operates completely with a push model, which means there's no central scheduler which defines test intervals, so test frequency is an emergent property, depending on how the admin has set things up. In simple cases it's simple, but
Client data is pushed up according to
-- John J. Boris, Sr.