Efficient Use of Disk Utilization Alerts

Like most we have a dynamic and widespread environment. For some time I have been able to get away with about 4 different disk volume utilization reports. They are:

1. Warning at 80%

2. Critical at 90%

3. Warning at 85%

4. Critical at 95%

(Each of the alerts has a corresponding custom property that is required to be set to 'Yes' on the target node)

- With more apps (servers) coming into our Orion instance I am finding that large disk drives do not necessarily benefit from the percent utilization conditions. As an example, a 2 Tb drive at 99% utilization has over 1980 Gb free. In our environment this would not constitute a critical type ticket.

- I know I am able to define disk utilization in alerts by Bytes, this however does not scale well. In theory each volume on each server could have different utilization threshold.

- I was trying to come up with some other conditions such as: Disk drive 99% used AND less than XX bytes. The problem with this approach is that I still need to define bytes that may or may not be applicable to the target device.

- We obviously don't have this issue with CPU and Mem as we can define at the node level.

- Does anyone have a method of disk utilization that scales in large environments ? I cannot have hundreds of disk utilization alerts as I fear it will negatively impact performance.

Find more posts tagged with

Accepted answers

All comments

orioncrack

The main problem is they do not allow granular overrides at the MB level, server level, group level.

This is a HUGE gap in the product in how disk alerting and monitoring is handled.

nick_scott

Orioncrack, that was a pretty optimistic post. Everything ok over there ? J/K, thanks for the feedback.

orioncrack

Alrighty then.

My main point is, with large volumes, % is not the way to manage.

We need MB level overrides.

if I sound cranky, I am because we have fought this same problem for years with SolarWinds.

With our other tool, we can override at the MB level, volume level. Easy money.

mesverrum

Volumes are able to do custom thresholds the same as CPU and Memory is done at the node level (although the default UI kind of hides the edit button for these, I use a custom query to add in a thresholds resource to my volume details page), but the threshold is unfortunately still % based so you wouldn't be able to use them to get around the issue of 99% being an acceptable value on a large disk.

I find that doing something like a volume percent used greater than critical threshold and space available less than 50gb or so works in most cases.

If I really needed to be able to tweak the minimum space threshold more granularly than the percentages + a global minimum, I would have to set up a volume custom property called minimumbytes and write my alert logic to make sure free space less than the volume customproperties minimumbytes. I would want to double check that the system handles nulls the way I expected though because I want to say that in SQL a null cannot be evaluated as a less than. Would probably need to set up an alert/trigger to automatically set all the minimumbytes to a global default value whenever it is null.

I've also seen cases where people just write volume space alerts with a series of OR condition groups so they can set various threshold based on the overall size of the volume, ie total size 2tb, critical threshold= 100gb, total >2tb<500gb threshold 50gb, etc etc.

Lots of ways to skin this cat, best is somewhat up to you and what you are comfortable with.

-Marc Netterfield

Loop1 Systems: SolarWinds Training and Professional Services

LinkedIN: Loop1 Systems
Facebook: Loop1 Systems
Twitter: @Loop1Systems

rschroeder

It's time to resubmit this to Thwack as a Feature Request / Idea.

It's time was due long ago.

m_roberts

It was interesting to see SW add a Volume forecast threshold value, but not a Utilised threshold value (Edit a volume and at the bottom you will see the threshold).

Volume threshold is the most obvious method, with the ability to select a value in % OR MB being suggested.

orioncrack

Indeed. It is.

That's how the other tool works. Easy peesy.

orioncrack

"I've also seen cases where people just write volume space alerts with a series of OR condition groups so they can set various threshold based on the overall size of the volume, ie total size 2tb, critical threshold= 100gb, total >2tb<500gb threshold 50gb, etc etc."

Me too, and its a nasty kludgey nightmare to manage when you're dealing with 6000 volumes.