Hey, everybody. My name's Gordon Cornelius here at Quest Software. Today I'm with Ben Boise. He's one of the solutions consultants here at Quest covering Foglight. I've been working with the product a little over 10 years.
And, Ben, I talk with a lot of customers and sometimes get them on the phone, and they just say, hey, what makes you different? What makes you special? What is the one thing that separates Foglight from the rest? And the very first thing that comes to my mind is performance investigator. Can you walk us through performance investigator and what makes it so special?
Yeah, absolutely. And the other thing that happens, and I've noticed it quite a bit, is inevitably these folks are getting either an email or a phone call or, heaven forbid, back in the days where we all used to work in an office, somebody walking up to you and going, hey, is there something going on with the database? And this is where SQL PI, I think, can really come into play.
Now, it's important to understand that Foglight is a very time-sensitive solution. And what I mean by that is not only can I look at things in near real time, but I also have the ability to go back in time. And Foglight's very good about providing certain predefined time ranges. We can see these listed here.
But I am, by no means, limited to just that predefined time range. If it is an eight-hour block, let's say, that I'd like to look at as an example, then I can slide back to a timeframe that's most important to me. Maybe something goofy was happening, if you will, between the hours of 12:00 to 7:00 AM and 8:57 AM this morning. I can drag that time frame, and as soon as I release, Foglight updates to show me data that was indicative of that particular time frame. And for folks that really need to get down to a nitty-gritty date/time range, we do provide very granular access to those investigations. I'm going to back this up and maybe do the last 24 hours, let's say.
And we're starting to see a couple of really updated charts. Now, at this point, there are a couple of things that I could consider. If there was something noteworthy that Foglight would show to me, I could see activity highlights. Hey, here are some things that we've noticed that might warrant further investigation. And it's kind of an easy button to lead me through that process.
And in my particular example, nothing really noteworthy has been brought to our attention, but that's OK. We have a lot of control over here. And that's another reality that I see with folks that are having to address performance concerns or provide information on what's going on with the database. They like this level of control. And so, now, it really is, what do I want to understand? And it may simply be, during this 24 hour period of time, what SQL statement was busiest in the environment?
And if I just expand that out, I can see that it's this SQL statement because it appears all the way to the top. But the reality is when I start to drill in a bit further into these individual dimensions, the bottom portion of the screen reflects to show me relevant metrics to what I've chosen.
So in this case, I'm getting some detail about the SQL statements, the top 25 SQL statements that we're looking at. If I were to choose that SQL statement, then I get very focused metrics on just that selected item. So through these dimensional investigations, I really get to roll up my sleeve and start asking questions about the workload itself.
Now, there's another approach to this. I could say, well, maybe I need to understand it by the resources that are being consumed. And I can see, once again, over the last 24 hours, that CPU usage-- most of the activity is related to CPU.
So if I choose that as a filter, I can start asking these same types of questions. What database or databases were consuming the most CPU during that timeframe? And the same sort of investigation, the same sort of information relevant to what I'm selecting here on the left-hand side is populated as I navigate through that tree.
In this case, I know right away this Quest [? Opti ?] [? new ?] database is consuming the most CPU over that 24-hour period of time. Well, let's go a little bit further. What SQL statements hitting that database during this time frame are appropriate or really root cause of the CPU utilization? So that's really one aspect of it, and it'll allow me to slice and dice, ask questions about the workload to be able to answer that question, hey, what's going on.
There's another dimension that's important to understand here, too, and that's something that we call advanced analytics. And so by default, there are different categories of changes that Foglight is keeping track of. And we can see these as checkboxes. So maybe I'm not really care what database objects, for example, have changed. But maybe I'm a little bit more concerned about the database configuration changes or the system configuration changes that have occurred in that particular environment.
So having this other dimension and what's changed that could be contributing to performance problems is very, very important. And similarly, Foglight's going to produce potentially some advisories. Hey, here are some things that we've observed during that timeframe. In fact, there have been about 18 occurrences of excessive I/O wait.
Well, tell me more. Show me not only what timeframe these advisories were generated, but also show me the relevant details-- in this case what SQL statements are contributing to I/O contention. Do I have tables that could be containing missing indexes? Give me some recommendations on things that I could-- steps, rather, that I could take to address these performance problems. So not only can I roll up my sleeves, but Foglight is doing a great job of assisting me through things like activity highlights, change tracking, and advisories.
And then, finally, what I like to talk about there is one last thing-- sorry, Gordon, this is a big topic-- is being able to prove or disprove that it's worse than before. Because sometimes folks will say, well, it seemed fine yesterday, but now it seems a little bit slow.
Well, maybe I want to prove or disprove that. I'm still looking at the last 24 hours. Maybe what I want to do is compare that overall activity to this same time yesterday or the week prior or to a specific date time. It's entirely up to me how I'd like to approach it, but then, through this comparison feature, being able to drill in and take a look at, OK, are things really worse now than they were previously. And in this case, I can see that this same timeframe a week ago, the instance was exponentially busier. It was a lot more active. It was consuming a lot more CPU.
And so maybe what I need to do is start to understand some of those same dimensions. Are there databases that are busier or less busy during that timeframe? Are there SQL statements that are busier or less busy during that timeframe? It's another dimension, if you will, another way to investigate performance problems, prove/disprove are things really worse than they were-- root cause analysis to understand where these problems are coming from, what these problems are, and then taking those corrective actions to address them.
Wow, Ben. That's a lot of information over a short amount of time. But I can see why people like the granularity and, again, the depth of information that Foglight can provide specifically in the Performance Investigator tab. Really appreciate your time today.
And those who are listening, if you're interested in learning more and having a deeper conversation in Performance Investigator, obviously there's a lot we can cover, so we're happy to have that conversation. Thank you, and have a great day.