Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials

Sustained Peak % Util?

I'm looking for a meathod to display a "top 20"-ish graph of interface utilization. In particular, I'm interested in seeing only interfaces which hit a sustained peak utilization of, say, 80% for a 5 min. or greater period of time. Ideally, I'd like to be able to generate alerts on such a condition, but I think a daily graph will suffice for the time being. Anyone have thoughts? Any sql guru's out there? =]

Thanks!
-Chris

Find more posts tagged with

Accepted answers

All comments

Jemini

your need a report of peak valve that greater than 80% for a while(such as today,or past two hours).
So you can chose custom report ,choose the interface whici has a peak value greater than a certain value.
Is it your want?

quote:Originally posted by bix

I'm looking for a meathod to display a "top 20"-ish graph of interface utilization. In particular, I'm interested in seeing only interfaces which hit a sustained peak utilization of, say, 80% for a 5 min. or greater period of time. Ideally, I'd like to be able to generate alerts on such a condition, but I think a daily graph will suffice for the time being. Anyone have thoughts? Any sql guru's out there? =]

Thanks!
-Chris

bix

That's sort of what I want, but not quite. I spoke with Orion support, and they said that it was outside the scope of their efforts, and that a custom SQL query might be the way to go. I talked with our SQL guys. Because of the fact that orion captures moments in time, what I'm really asking or is an application to ask something like: If a given record for interface utilization is above 80%, then make note of it and look at the next record. If that record is also above 80%, then make note of it and look at the next record, etc. If the >80% utilization continues for more than 5 mins worth of polling records, make note of how long the duration lasted. If the >80% utilization continues for less than 5 mins, then ignore the spike and move on to the next record.

Jemini

quote:Originally posted by bix

That's sort of what I want, but not quite. I spoke with Orion support, and they said that it was outside the scope of their efforts, and that a custom SQL query might be the way to go. I talked with our SQL guys. Because of the fact that orion captures moments in time, what I'm really asking or is an application to ask something like: If a given record for interface utilization is above 80%, then make note of it and look at the next record. If that record is also above 80%, then make note of it and look at the next record, etc. If the >80% utilization continues for more than 5 mins worth of polling records, make note of how long the duration lasted. If the >80% utilization continues for less than 5 mins, then ignore the spike and move on to the next record.

I think your request could be achinved ,anyway. It's all about sql task,but look at the poller ,which poll data from nodes or interfaces
every certain time(such as 5mins),but not all the the elements were polles at the same minute and the same second.so that you need study
the database and then make your sql query.
I'll try and post the code out later.

Network_Guru

On a large Orion SLX installation you might find polling your interfaces at 5 minute intervals is excessive. I only poll my backbone interfaces (a handful at best) every 5 minutes. All others are polled at 15 minute intervals.
In this case you need a sustained utilization over 15 minutes to set off an alert, which is basically what you are looking for? (5+5+5=15minutes).
This will tend to even out any short bursts, but let you know when the lionk is pegged for 15 minutes or more.

-=Cheers=-
NG

iceman

Hi Bix,

Did you ever resolve this? I'm also intrested in this. One of the main questions I receive is how often I hit a particular %utilization.

Thanks

bmoore_222

I am also looking for a way to alert/report on average sustained utilization

Jemini- Did you have any success with your SQL query?

Has anyone successfully created this top 20-ish graph on sustained interface utilization and/or an alert on it?

Network_Guru - How can you set off an alert on sustained utilization like you are mentioned below? When setting up alerts I only see average transmit and average receive traffic, but no sustained or for X amount of polls.

Network_Guru

The alert is based on the utilization over the duration of the statistics polling interval.
If you poll for the interface stats every 10 minutes, then you must exceed the threshold for the full 10 minutes before you get alerted.

-=Cheers=-
NG

clisser

quote:Originally posted by Network_Guru

The alert is based on the utilization over the duration of the statistics polling interval.
If you poll for the interface stats every 10 minutes, then you must exceed the threshold for the full 10 minutes before you get alerted.

-=Cheers=-
NG

Are you sure on this? It is my understanding that Orion will poll the interface at the 10 minute intervals (a snapshot) and if your interface is over the threshold at the time of the poll it will trigger the alert.

Now what would be ideal would be to have the option of alerting if an interface was over the threshold for x number of polls.

Network_Guru

The statistcs collection method is explained in the Orion documentation.

It would be a great feature to alert based on percentage change. This could be used for any kind of alert, such as utilization, packet loss, latency, cpu, etc.
If your CPU normally runs at 15% & suddenly spikes up to 40%, then send a notification.
It would be based on percentage of change, and not on an actual number, so that it could be applied universally to all devices, without having to go and create a custom alert for each one!

-=Cheers=-
NG