-
Re: Polling at different intervals?
dodo123Oct 20, 2017 3:53 PM (in response to Antonio)
think you can set multiple node polling from manage nodes select them and edit them all change the polling frequency.
so live 5mins
preprod 10min
test 15min
But i don't think you can tell them when to poll as you could probably bring the system to a halt if you could.
Why do you want to do this?
-
Re: Polling at different intervals?
Antonio Oct 20, 2017 4:02 PM (in response to dodo123)Cause I have a large flat network where the switches (L2) are not able to manage all the polling at once and eventually their buffers start dropping packages. without going into details, I can not afford packet drops at all since every packet is of the most importance.
-
Re: Polling at different intervals?
dodo123Oct 20, 2017 4:15 PM (in response to Antonio)
No worries so you want your polling spread out?
the system should be handling that on its own if you polll all the devices at the same time wouldn’t that be worse or am I getting the wrong end of the stick.
-
Re: Polling at different intervals?
Antonio Oct 23, 2017 7:46 AM (in response to dodo123)Hmm, my understanding has always been that polling is done at once system wide. I'm not aware that it is spread out, which is what i'm trying to do.
-
Re: Polling at different intervals?
dodo123Oct 23, 2017 8:04 AM (in response to Antonio)
I would be amazed if it was that would explain a lot of the problems I am having.
and if you have multiple nodes open you can see that they poll at different times but there will be some that have the same time but im sure its a queuing system
as when you click poll now its not actually now it just adds it to the end of the current list of polling devices.
-
Re: Polling at different intervals?
Antonio Oct 23, 2017 9:38 AM (in response to dodo123)I just double check and most of the devices poll either at 180 or 240 sec (default values) but they are grouped by device type. I will have to breakdown all those switches into different time frames to solve my issue.
-
Re: Polling at different intervals?
dodo123Oct 23, 2017 10:22 AM (in response to Antonio)
the interval time will be the same across most devices but that doesnt mean all the device will get polled at 14:00, 14:05, 14:10... it means batches will fire at different times and the polling interval will be the same
server A added or forced polled at 14:00 this will poll every 5 mins from when it was added. 14:00, 14:05, 14:10..
server B added or foced polled at 14:01 this will poll every 5 mins from when it was added. 14:01, 14:06, 14:11..
server C added or forced polled at 14:02 this will poll every 1 min from when it was added. 14:02, 14:03, 14:04, 14:05
Hopefully someone can confirm this but im sure it queues it and doesnt just fire everything at the same time
-
Re: Polling at different intervals?
mesverrumOct 23, 2017 7:36 PM (in response to dodo123)
You are correct, and it is granular down to the ms. You can see by looking at the timestamps for just about any data point in the db, I pasted below an example from my lab showing from one poll to the next for a bunch of devices on a 10 minute interval, you can see they are mostly staggered every few seconds, but sometimes clump up. The scheduler does the polling on a rolling list based on the time a device was added + the polling interval, doing a poll now will insert an extra datapoint into the schedule but does not change your normal scheduled slot. I think the only way to get your object reassigned in the scheduler would be to unmanage and remanage the node.
70 2017-10-23 23:39:56.503 1 2017-10-23 23:39:54.593 76 2017-10-23 23:39:48.27 20 2017-10-23 23:39:20.297 44 2017-10-23 23:39:18.373 9 2017-10-23 23:39:07.437 89 2017-10-23 23:39:05.67 58 2017-10-23 23:39:02.407 55 2017-10-23 23:38:59.517 69 2017-10-23 23:38:28.907 88 2017-10-23 23:38:13.273 67 2017-10-23 23:38:11.057 189 2017-10-23 23:37:05.79 191 2017-10-23 23:37:05.79 165 2017-10-23 23:37:00.447 178 2017-10-23 23:37:00.447 179 2017-10-23 23:37:00.447 180 2017-10-23 23:37:00.447 182 2017-10-23 23:37:00.447 190 2017-10-23 23:37:00.447 188 2017-10-23 23:37:00.4 166 2017-10-23 23:36:56.977 184 2017-10-23 23:36:56.977 41 2017-10-23 23:36:56.367 84 2017-10-23 23:36:56.367 86 2017-10-23 23:36:56.367 82 2017-10-23 23:36:55.353 22 2017-10-23 23:36:52.21 43 2017-10-23 23:36:52.21 80 2017-10-23 23:36:52.21 85 2017-10-23 23:36:50.883 19 2017-10-23 23:36:50.727 24 2017-10-23 23:36:50.727 6 2017-10-23 23:36:50.383 30 2017-10-23 23:36:50.383 46 2017-10-23 23:36:50.383 47 2017-10-23 23:36:50.383 48 2017-10-23 23:36:50.383 161 2017-10-23 23:36:50.383 183 2017-10-23 23:36:50.023 192 2017-10-23 23:36:50.023 193 2017-10-23 23:36:50.023 169 2017-10-23 23:36:49.837 163 2017-10-23 23:36:49.633 53 2017-10-23 23:36:49.477 49 2017-10-23 23:36:49.367 62 2017-10-23 23:36:48.163 181 2017-10-23 23:36:40.383 29 2017-10-23 23:36:27.037 103 2017-10-23 23:36:23.507 20 2017-10-23 23:36:21.553 35 2017-10-23 23:36:16.35 34 2017-10-23 23:36:14.63 196 2017-10-23 23:36:12.99 31 2017-10-23 23:36:04.787 111 2017-10-23 23:35:58.333 79 2017-10-23 23:35:48.583 68 2017-10-23 23:35:39.17 73 2017-10-23 23:34:56.237 74 2017-10-23 23:34:22.683 20 2017-10-23 23:34:20.28 44 2017-10-23 23:34:18.36 89 2017-10-23 23:34:05.657 81 2017-10-23 23:33:13.243 103 2017-10-23 23:31:23.493 78 2017-10-23 23:30:55.423 70 2017-10-23 23:29:56.5 Do you show a lot of packet loss on your nodes Antonio? Orion only sends out a single ping for each up/down poll so if your buffers were being crushed on a regular basis it would definitely show up there. I honestly would be pretty worried about any network brittle enough that Orion is over running the switches. In my lab environment running every module with 500 elements we are maxing out at 3 mbps of traffic (including writing the polled data back to the sql server and serving up pages to myself and the other engineers using the lab each day. If I project that out it would seem that maxing out a primary app server polling 12,000 elements and a stack of SAM monitors and VMAN polling and such would probably move about 70-100 mbps of traffic to monitor thousands of switch ports and servers.
You can see Cisco buffer errors if you look at the network tab on the node details view. Click on the blue text to see the errors charted out over time. Check this on the nearest switch to the orion poller and if your problems stem from polling then you should expect to see a nearly continuous stream of misses and if things are really bad then there should also be errors (a miss isn't a dropped packet until it becomes an error)
Hopefully your network devices support QoS so you can prioritize the traffic getting dropped because honestly if you are worried that your monitoring is tipping you over the edge of the port capacity then you will be really jammed up if anyone tried to copy a large file from the file server or print a pdf to a network printer.
-
Re: Polling at different intervals?
dodo123Oct 24, 2017 3:24 AM (in response to mesverrum)
Thanks for the detail, I was beginning to doubt my own thoughts
-
Re: Polling at different intervals?
Antonio Oct 25, 2017 11:48 AM (in response to dodo123)Dodo, You and mesverrum validated my thinking. Thanks. I believe that I failed to clarify one very important thing. The network in question is not a Cisco network, the switches are Siemens (L2) switches. I would definitely not have this problem if they were. Anyways, just so you know. I have two completed isolated networks of different size, with their own NPM server. the smaller one does not shows any signs of degradation while the bigger does.
-
-
-
-
-
-
-
-
-
-
Re: Polling at different intervals?
dodo123Oct 20, 2017 4:20 PM (in response to Antonio)
Additional polling engines may help also using agents everywhere.
Stay away from wmi I guess.