Hi,

The behavior desired in item #1 can be achieved by using a "Stateful" Comparator.  We are working on a similar threshold based alert for ESX server and here is a quick snapshot of it -

In this case, the Comparators being stateful (keep the Stateless flag unchecked), will only generate the SNMP trap when the condition is met the first time.

For item #2, you could use the mail action in addition to the SNMP trap action already in place.  Alternatively, Alert Consolidation Email Notification (located in the Alert definitions node) can also be used.

I hope this helps.

Thanks,

-Gayatri

hi Gayatri:

thanks you to give me a different hints.

As my example, if the CPU utilization is  over 75% and last for 2 hours, how many alert notification will be created in your definition ?

example 2: if the CPU utilization is  over 75% for about every 10 mins? how many alert notification will be created in your definition?

Hello Power (do pardon me if that's not your first name),

With the alert definition I provided, only 1 alert will be created in both your examples.  The thing to note though is that this alert will remain active with Severity getting updated until the data comes in suggesting that CPU Utilization has gone below 75%, which is when this alert gets cleared.  So, now when you consider the Alerts Dashboard, an active alert indicates a situation (in this case High CPU Utilization) exists, which gets cleared only when the situation's resolved.

Let me know if you had something else in mind...

-Gayatri

hi Gayatri:

just call me power...

Your thought is good too, but I have different purpose with yours.

We will not watch dashboard all the time, but monitor the utilization by email.

The purpose I want is that sending a email every 60 mins when the average utilization during this period is over threshold.

in my example:

1.if the CPU utilization is  over 75% and last for 2 hours: I will get 2 email alert

2: if the CPU utilization is  over 75% for about every 10 mins, i will get email alert when the average value is over 75% during last 60 mins; if not, i will not receive an email.

This configuration works fine in watch4net 6.2, but not in SRM 3.5, I don''t know what's the difference.

Ah, I see.  Let me give this a quick try and get back to you, Power.

Hi Power,

Apologies for the delay - we had some network issues yesterday.

So, I gave it a try and it worked for me.  Here is the alert definition -

And the Time Window definition looks something like this -

Is this what you have tried and has not worked?

-Gayatri

yes, that's the configuration.

my focus is array controller utilization, and my time windows setting is 60 mins, threshold setting is more than 70%. it always send out mail every 20 mins not 60 mins, so strange.

for example:

SRM3.5

W4N6.2

Worked for me.  Thanks Gayatri!

Absolutely Power!  We are with you all the way.  Do let us know either way...