cancel
Showing results for 
Search instead for 
Did you mean: 
Create Post
Level 12

Settle an argument between me and my boss.

Hi Thwackers.

My boss and I had a discussion today and I need your help to settle an argument.  We recently had to beef up our server and our polling engine due to it being sluggish.  My boss is saying that we have to lessen the scope of the alerts that go out, 

Here is what I mean by this.

My thought is that if you create a SAM monitoring (see screen shot #1).

pastedImage_0.png

Then an alert is created to email out to the group.   here is how it was created.

pastedImage_2.png

So does this mean that the scope of the alert searches ALL nodes looking for that service? 

or

am I right that it is only looking for the nodes it was assigned to?

or

Neither

or

Both somewhat right?

Thanks in advance

jeff

0 Kudos
4 Replies
Level 12

sean.martinez​ I have to create 6-8 different custom properties since that is approx.  The number of different email address I will be sending to depending on the node and what the issue is?

0 Kudos

trilobite rex​, I believe sean.martinez​ is referring to is a single node custom property, where you would enter the email address for the service owner of the application running on that server (i,e, whoever administers the application). It would be a single custom property creation, but the data would change depending on who is responsible.

i.e. your email servers would have the email address of your email server admin in the node custom property "Point_of_Contact", using sean.martinez​'s example.

Using the information in the link Sean also provided, you can automate the alter trigger email notifications to go to this email address using the variable $(Point_of_Contact) as the data in the 'Email To' field.

- Jez Marsh
0 Kudos
Level 16

If Orion is slow then look at your database server -- what is it doing?

Alert evaluation and actions are only executed on the main application server, so beefing up a polling engine will not improve performance there.

The alert evaluations should be hitting indexes and recent data in the database -- your sqlserver should not be hitting the disk for this type of query: does it have enough RAM to effectively cache/buffer your data, or is it stuck in diskio because you don't have enough RAM there.

Look at Buffer Cache Hit Ratio in the Performance Monitor under “SQLServer:Buffer Manager” -- mine is running at > 99.9%

0 Kudos
Level 14

The Alert currently configured will only check when the entire Application name is iMFmaSvc, it needs to be contains for it to function.

The best practice is that you should try to setup as few Alerts as possible to encompass as many Applications as possible. These alerts get ran against the database every minute. Lets say you have 200 applications, if each application had an alert, 200 alerts checking every minute, 1440 minutes a day, so that means the database is scanned 288,000 times a day. Reducing the number of Alerts to use Custom Properties or Warning and Critical levels will reduce the need of similar alerts needed to be created, and reduce the impact on the MS SQL Server database.

What you can do is remove the Application Name, and just use the Status of the Node and Application. What I would then do is create a Custom Property Field (Point_of_Contact) and enter the Application or Nodes POC's email address to that field, this way the To field would be ${Point_of_Contact} that would auto-fill the email address. This document covers it in more detail Using Custom Properties sending Alert emails