On 16-08-2013 21:30, John D. Alexander wrote:
A couple of questions about Xymon capabilities…
-Can Xymon handle hosts that cannot be contacted by the server? I will have hosts reporting in, but won’t know the IP and cannot connect to them. Status reports will actually come from a central agent rather than from the devices
Works fine. As long as your central agent sends data using the Xymon syntax of "status HOSTNAME.TESTNAME ..." then you can send data on any server.
There are no checks in Xymon on the source of a status message. You can enable an IP check with the --status-senders option so you probably do *not* want to enable this.
If you cannot ping the hosts from Xymon, add the "noconn" option in hosts.cfg to disable the ping check.
-Can Xymon handle multiple different polling periods – specifically with RRD? Some devices check in every minute. Some check in every 2 minutes. Some every 5 minutes.
Short answer: It will work, as long as you send updates at least every 10 minutes. The RRD files will just discard the extra updates.
The longer answer: RRD files have two settings telling them how often updates arrive - 'heartbeat' and 'step' - and Xymon sets these to 10 and 5 minutes, respectively. Which means that you *must* update the RRD at least every 10 minutes for it to regard any of the data as valid; and if you don't update exactly every 5 minutes, then RRD will extrapolate your data and compute what the value would have been if it had arrived exactly on the 5 minute mark.
You can configure these - on a "per-status-name" basis - in the rrddefinitions.cfg file (although I've had some reports that it may be buggy). But that is for defining a specific set of RRD files that have different polling periods - not for some hosts reporting the same type of data at different intervals than others.
It really is an RRD thing, so I would recommend that you look at the "rrdcreate" man-page - especially the section on "The HEARTBEAT and the STEP" - if things don't work as expected.
Regards, Henrik