Hi, and welcome to this introduction to Spotlight on SQL Server Version 11. My name is Peter O'Connell. I'm the Product Manager for Spotlight on SQL Server Enterprise.
Before we begin, just take note of where you can get Version 11. If you go to Spotlight on SQL Server Enterprise, that will get you to the download page for Version 11. Also, check out our New Resource Center. A lot of new resources going out there and videos that will follow on from this one to better inform you on how to get the most value from your Spotlight investment.
There are some key things which we have implemented within spotlight over our recent roadmap. One is Simplicity Without Compromise. To that end, we've implemented two key changes here. First of all, we have a New Look User Interface. This gives a Windows metro look to the existing UI. Fans of the UI will be glad to know that all of the key main workflows remain intact and without change.
Really what we've done is we've modernized the look and we've modernized the components within the UI, which will give us more flexibility in future releases to do more exciting stuff. Also the Planned Outage feature, which we introduced in Version 10.5, we've taken your suggestions on board in terms of improvement and implemented quite a few of them. So let's take a quick look at that now.
Looking into Spotlight Version 11 you see that nice Windows Metro look and feel. Couple of improvements to the heat map. You'll see the fact that the connection type is noted on each of the titles there, but essentially it remains the same flow as before. Also you'll see reports are tidier. They've got a tab of their own now. And are broken into views, reports, and web reports.
If we go to the Configure Tab and look at Planned Outage now and go to Add a New Planned Outage Rule you'll see for a start we've implemented a suggestion that everybody came up with almost immediately, which was can we implement this on multi-select instances when we're creating a room. So that's there.
And also under monthly you'll find that there's a lot more options there in terms of re-occurrence on a monthly basis. So moving back to our key themes, the next key one for us is getting the full picture. And again, we've always attempted to give the full SQL Server stack. The last Version 10 we introduced Wait Stat Analysis. In a similar mode, we've introduced the same kind of multidimensional breakdown for workload analysis in Version 11. So let's take a quick look.
We go to our 02 instance here. Take a look in. And you'll be familiar with our Wait Events screen from the last release. So here we see a summary of the Wait Events over the last hour. And it can drill down by database. So here we get a list of the databases, the biggest consumers. In fact, if I go for, say, the last two hours or so, order will change slightly. That's thanks to the first one, PARALLELISMTEST. PARALLELISMTEST is basically the script on Paul Rendell's e-learning module up on [INAUDIBLE] site. Well worth checking out.
So there's the sample database, PARALLELISMTEST, which is number one there. I can click into PARALLELISMTEST, drill down into that. And because this was a demo around parallelism, you see that, no surprise, that CXPACKET is up there as number one. Right.
So Wait Events initially introduced the effect or the symptom of what was happening within the system. Then we added another dimension, which is called Workload Analysis. And in terms of Workload Analysis you can see use that same multi-dimensional drill down right across, CPU, Duration, Logical Reads, Physical Reads and Writes, and so forth.
So if I go to Database again, let's see the order that exists right now. And go for the last two hours. We'll see my order reordering again, and again PARALLELISMTEST comes right to the top there in terms of the load test we were running on that earlier on. We can click into that again and see straight away the percentage of CPU that was consuming. I can look at the applications. So you'll see there are SQL Server Management Studio.
Also there's our Diagnostic Server. And just to show that our overhead is less than 1%, you can see here this is down it, I think, it's 0.01% or something. But, of course, the bulk of it coming from those scripts that we're running. And we can drill right into those scripts to see what exactly was the script that was running at that point in time.
So you can do that and dissect your workload by user, by host. So rather than just focusing on the effect or the symptom, we now focus on the cause and we use all the other diagnostic drill downs within spotlight to give us really extensive and fast root cause analysis.
So the other piece we've introduced in terms of getting the full picture is Baselines. We can take any counter that is graphed within Spotlight and we can compare against a Baseline. In this case my offset is 24 hours. I have 6 intervals. So that's 6 by 24 hour is the data set that it's using to give me a statistical representation here.
And then I can see whether this particular metric, and this particular point in time, or this peak is within the median, whether it's in an interquartile, outerquartile, and basically whether this is normal behavior are not.
So that brings us to our next thing, which is Tool of Choice. So a lot of you out there are implementing systems such as SCOM or have other third-party systems. You might have event management systems and so forth that you want to integrate with. And you wanted a mechanism to do that within Spotlight. You'll
08:53