Hello, my name is Jason Hall. And today we're going to be taking you through a feature of Spotlight on SQL Server Enterprise called alarm actions. Alarm actions are a new feature in version 10.0 of Spotlight that essentially decouples the action from the alarm itself. Alarm actions provides an Outlook-style rule designer that allows you to simply yet powerfully configure events that you wish to take place in response to alarms.
Several of the use cases for alarm actions are now listed on the slide. And we'll be going through some of these in the demo. But to talk through some of the more common scenarios, alarm actions are going to allow you to take different actions on different alarm severities. Perhaps you want to send an email to the DBA group at a medium severity alarm, and then to the network operations staff at high.
Another use case is you want to take the same action against a group of alarms. Maybe for your 10 most critical alarms you want to write them to the Windows Event Log. Another common scenario is when you want to apply the same action to a group of servers. Perhaps you have two DBA teams. And you want to send an email for your East Coast servers to DBA team one and you want to send an email to your West Coast servers to the second DBA team.
Another scenario we hear quite commonly is that you only want to take action on an event during specific days and times. Perhaps you only care about CPU problems during your business hours. And, lastly, maybe you want to take an action on alarm only if it's been active for a certain number of minutes or hours. Perhaps you don't care about spikes in page life expectancy, but if it lasts for more than 10 minutes, you want to take an action on it. Let's go ahead and see how this feature works.
So here we have Spotlight on SQL Server version 10. And in prior versions of Spotlight on SQL Server, you may remember that the alarm itself had a set of actions that you can configure in the lower right-hand corner. You could send an email or run a program. In Spotlight version 10 and higher, you'll no longer see the actions in the alarm because we've effectively decoupled it into a separate interface called alarm actions.
Now I'll walk you through the creation of one of these alarm actions. And then I'll show you some examples of some additional options that you have. So let's say, for example, we want to create an action that tracks SQL Server availability, but only for our West Coast data center.
So we first name the action. And then we define a condition. Now for this purpose, I first want to pick the set of alarms that I want this action to take place over. So we'll check off The alarm is.
And then down at the bottom of the screen, we're going to pick the specific alarms that we want to include. For this case, we're tracking outages or availability problems. We'll scroll down and find my SQL Server Connection Failure alarm. And then I'll find the complementary alarm down here on the Windows side-- so both SQL Server and Windows Connection Failure.
Now that we have our alarm selected, I'll come in here. And this only should apply to our West Coast servers. So I'm going to go down here to the connection tag condition.
Spotlight allows you to tag a server using a specific property. And in my case, I've tagged my West Coast servers using this DataCenter West tag. So I'll add that tag in now as an alarm action condition.
And then, finally, maybe we only care about availability issues on Monday through Sunday. We do reboots on Saturdays. So we're not concerned about availability for this set of servers unless the problem happens on a day Sunday through Friday.
Then, finally, we have the ability to set up our actions. We've already set up our condition. And we can track the description down here the bottom.
So let's say now I want to send an email based off of that condition, too. We've got a distribution list set up. So we'll just call this West Coast DBA Team at dot com.
Now you can also see some additional actions that we now allow for in version 10 of Spotlight-- executing SQL scripts, executing PowerShell scripts, writing the alarm details to the Windows Event Log. So we'll do simple email here. But you do have plenty of options. And that's really as simple as it is to build that fairly complex alarm action based off of sending an availability email to a specific DBA team only for a specific set of servers.
Now some of those other use cases I've already built here. So let's go ahead and take a look at the production alarm during business hours. You can see, again, I've set up the condition. So for these particular performance alarms, I'm tracking I/O Wait Time, Blocked Processes, CPU Overhead, and VMWare Overhead.
When the server tag is a production server-- that's another tag that I've set up-- in this case, I want this to track Monday through Friday. And then I determine that the business hours for this group of servers are from 6 AM to 7 PM. And then, in this case, I want to write that event to the Windows Event Log. And that's actually a very useful action that you can take place when you possibly want to integrate Spotlight's alarms into some other system that can scan through the event log.
And then the last use case I'll show you is that duration-based alarm action. Now in this case, we only want this to apply to a single alarm. And that's page life expectancy. Again, the condition here is for my production servers.
06:21