Closed

Closed due to inactivity. Received 145 votes with last vote on 11 Sep 2019.

Node Specific Thresholds - v2.0

I know this has been thought of and there are even existing feature requests, and may even be in the works, but I feel this needs more attention, and perhaps my level of detail will help.

Having a global settings (default) is great, but we need the capability to fine tune the node to specific settings for each hard drive and memory resource.

The easiest way that I could think to do this, is by adding a Thresholds button to the Management panel of the node.

threshold.png

With that button the user is prompted to Enable the Node Specific Settings

Thresholds Dialog.png

After the setting has been Enabled the user can go to the Disk Volumes and modify the default thresholds for the specific object.

Threshold Settings.png

In the Node Thresholds section the user will then be able to modify the thresholds for CPU, Disk, Memory and Response Time.  An additional option for Disk thresholds will be added allowing the user to specify (in MB) the minimum amount free before alerting.  This would be available for EACH object, and can be different between each object.  The option of % Used versus Amount of MB Free is determined by the checkmark next to each option.  If one is checked the other is unchecked.  If neither are checked the default of % Used will apply..

Thresholds Dialog 2.png


Here are some of the other Feature Requests, please be sure to vote for them as well.

http://thwack.solarwinds.com/ideas/1846

http://thwack.solarwinds.com/ideas/1623

http://thwack.solarwinds.com/ideas/1678


  • I have implemented something very similar using volume custom properties, v_warnthreshold, v_warnthresholdGB,  v_critthreshold, v_critthresholdGB.

    I then recreated the disk alerts, using SQL to accept the threshold that is populated and alert upon it.

    This allows us to alert when there is x GB left rather than relying on (integer) percentages.

    However none of the volume widgets are capable of displaying anything but percentage values. So the alerts work great, but volume view does not reflect this.

    So bump!

  • GREAT idea....  big time saver!!

  • This would be great.

    Ability to put upper and lower thresholds would be helpful also.

  • Does anyone know if it is possible to reverse the threshold? When I edit a volume it is by default set to greater than. I can override the threshold, however not change greater than to less then.

    We have some devices where it should be red if the usage is below – let’s say 50%, then the cache is not working. I can do that with an alert, but would like the chart to stop being red as it is not a problem that it is using 99%.

  • FWIW I assume physical/virtual memory is under volumes, because that's where windows puts it in the snmp mibs.  The memory gets index IDs in the same OIDs as disks.

  • It would be nice if the managed nodes page had different "views".  Quickly pulling up a dropdown and selecting would show you all the settings for a specific set of features or settings.

  • NPM 11.5 RC2 has individual thresholds for Volumes.  It doesn't look quite as pretty as what cczech suggested, but its there.  You can also configure your alerts to use Node threshold values as well instead of static numbers, but that feature has been available as long as I've known it, even for volumes.  Its just before you could only configure volume thresholds globally and now you can do them at the volume level.

    Go into Manage Nodes, hit the plus button beside the Node(s) whose Volume(s) you want to set thresholds on, select the Volume(s), then click Edit Properties.  There you can set a percent threshold on the Volume(s) you selected.

    My only gripe is that if you want to change a lot of volumes' thresholds at once then it takes a LOT of hunting and clicking (unless I'm missing some other method that I'm not aware of...).  Say for example you want to set all of your C:\ Drives to have a 97% Critical Threshold.  You have to expand every single Windows Node by hand and select its C:\ Drive...  If you only have a couple Windows servers, no big deal, if you have a couple hundred (or even 10 or 20 really...) then you will be expanding and clicking for a LONG time.  The really need to give Volumes their own drop down box in the Manage Nodes screen just like Interfaces do.  That way you could just search for C:\ and then select all, click Edit Properties and be done...

  • what happen with this request? hoping that, this will be available soon emoticons_happy.png

  • I might be misunderstanding, but are you saying that you want individual volume thresholds?  I ask because you also mention wanting CPU, RAM, and Response Time thresholds on a per Node basis, which has been a feature of NPM for some time now.  Like I said, I'm sure it is just me not understanding what you are asking here since you have a lot of people agreeing with you, so I just wanted to ask for more clarity I guess.

    EDIT: Alright, I re-read the beginning of your post again and I see that you did specifically say that you would like thresholds on individual hard drives and memory resources.  I know this feature request was written before this feature was announced, but for those reading this post now and don't know it, NPM 12 does have the individual Volume thresholds feature and is pretty far along in the beta process, so hopefully it should be GA in the next couple of months (hopefully, I really don't know when it will be out, I'm just guessing based on past releases).  As far as the thresholds for individual physical and virtual memory, I'm assuming the new individual volume threshold will extend to those since Solarwinds treats those as Volumes for some reason (I'm not a fan of that, it should be separate or, at minimum, have the physical and virtual memory polling be built into the CPU & Memory poller instead of the Volume pollers.  If you want to monitor your physical and virtual memory monitored separately then having it read as a Volume means having to manually filter all of the Volumes related resources on views to omit those volume types. 

  • Seriously needed....

    Great request