UDT Ports Auto Unmonitor

Question

We have an issue where if a particular switch goes offline (usually due to an ISP at a remote site, from a power outage, or sometimes even with the switch rebooting) the UDT system automatically marks those ports on that device as unmonitored.  When the device is polling again UDT doesn't automatically restore those ports as monitored.  Usually I will come across a switch and notice that the port information isn't appearing like it should and I will check the ports and they went unmonitored again.  Its easy enough to re-activate the monitoring.  It would be nice though if it didn't do this.  Is there a setting that manages this?  Is it based on a time frame of being offline?  Is there a way to send an alert when the UDT automatically unmonitors a port?

rogdog · Accepted Answer

Hi lazyrabbit,

Solarwinds support confirmed the following to me

Why do ports become unmonitored ?

If a port/switch is unreachable in UDT for more than 3 Hours, it goes into an unmonitored state.
(we have feature request open to extend this time )

Cause
This is by design!

There is a new option in the UDT Database.
UDT settings table
UDT.UnmonitorPortDelayHoursWhenNodeIsDownOrReboot
By default, this is 3 hours.
If an end device is not reachable for this period, ports on this device will be shown as unmonitored.

Resolution

You can manually increase this period. For example, to set it to 10 hours:

UPDATE UDT_SETTING SET SettingValue=10 WHERE SettingName='UDT.UnmonitorPortDelayHoursWhenNodeIsDownOrReboot'

"Setting UDT.UnmonitorPortDelayHoursWhenNodeIsDownOrReboot to '0' will mean that the ports are not monitored.
You can set the setting to a maximum of 256204776 hours, this equates to over 29,247 years"

There is a longer discussion on this topic here How might I change UDT so switches' unmonitored ports are monitored?

DATE: 16th November 2018

WARNING: ***** Do not set value to 256204776 **** Unknowingly to me, setting the value to 256204776 actually broke my UDT. Everything looked fine but ports weren't being updated. After opening a case with Solarwinds they asked me to change the setting to 87600 (10 years). This seem to have fixed the issue.