Monitoring SLAs with SQL Monitor Reporting

Proactive Reporting for SQL Server

If you’re a return reader of this blog you know I write often about monitoring and performance of Availability Groups. I’m a very big proponent of using monitoring techniques to ensure you’re meeting your service level agreements in terms of recovery time objective and recovery point objective. In my in person training sessions on “Performance Monitoring AlwaysOn Availability Groups”, I emphasize the need for knowing what your system’s baseline for healthy replication and knowing when your system deviates from that baseline. From a monitoring perspective, there are really two key concepts here I want to dig into…reactive monitoring and proactive monitoring.

Reactive Monitoring

Reactive monitoring is configuring a metric, setting thresholds for alerting and reacting when you get the alert. This type of monitoring is critical to the operations of your system. The alerts we configure should model the healthy state of our system…when our system deviates outside of that state, we certainly want to know about that so that we can act…well really react accordingly.

Proactive Monitoring

Proactive monitoring with an alert based monitoring tool is a little harder. What DBAs and architects do is periodically sit down and go through their existing monitoring systems and review the data over some time interval. And if we’re honest with ourselves we try to do this at regular intervals but don’t get to it very often because we’re busy. And when we do finally get in there to look it’s usually to do a post mortem on some sort of production issue…then very often we find that a critical performance attribute had been slowly creeping up over time until it reached a tipping point and caused a production issue. We do our analysis, make our system corrections and move on. Still not exactly proactive. Mostly because there is still a person in the process.

Reporting on System State

With our Reactive Monitoring model, we already define what a health system state is. And let’s take people out of the equation. In Redgate’s latest release of SQL Monitor they added a reporting module. In here you can define reports that will represent the state of your system and you can get a snapshot view of what’s critical to you about your SQL Server environment. So if you are running Availability Groups, like I mentioned above, you can configure your report to have those critical performance metrics already set up so you can quickly get to them via the SQL Monitor interface. But better yet, you can schedule the report to be delivered right to your inbox. Yes, another email. But the report is a simple PDF…you give it a glance, process the data and move on. This saves you from having to go into SQL Monitor’s web interface to view the state of your system. Find something odd, jump into the Web UI and start digging for the root cause.

Reporting gives us two key advantages

1. Point in time snapshots of our system’s health – we’ll get that report dropped into our mailbox and on that report is the most critical performance metrics that define the health of our system.
2. **Ability to see slowly changing trends **– reporting helps us focus on trends. It allows us to zoom out a bit and view our system over time. Less reacting, more “proacting”

OK and one more…for me, as a consultant, I can define reports for clients and have them emailed to me periodically to review. Let’s take a second to build a simple report together

Creating Your Own Reports

Now normally, I’d show you a bunch of screenshots on how to create a report in SQL Monitor…but rather than do that…go head and click on the menu below and start trying out the reporting interface yourself using Redgate’s publicly available SQL Monitor Demo Site! 

A couple reports I think you should look at right away are

1. Example Report – this is the landing page on the demo site and as you can see we get a snapshot of our servers’ performance.
2. SCC – Custom Metrics – in SQL Monitor you can add your own custom metrics, these are things that you think are business critical…the report in the demo site shows you SQLServerCentral custom metrics of Emails Sent per Hour and Forum Posts per Hour.
3. Storage – here you’ll find things like fastest filling disk, disk space used, and database sizes. 
4. Create Your Own – Download the trial version and create your own report. What are the things that are important to you that you want to have right at your fingertips when you log in or have a report land in your inbox with that data?