
It’s 3 a.m. and suddenly your phone blows up with alerts. Something really bad must have happened to the server. Only it didn’t. It’s another false alarm—the third this month, in fact. As you head back to bed shaking your sleep-deprived head, you make a mental note to review the system alert settings tomorrow and make some sorely needed adjustments.
The scenario above illustrates what happens when alerts and alarms aren’t set up efficiently. Alarm fatigue sets in, which increases the chance of a DBA missing a critical alarm. Alarm fatigue occurs when a DBA is constantly barraged by alerts—some important, most not.
Fortunately, you can customize your SQL Server alerts so you only receive those that need prompt, human attention.
The easiest way to calm the chaos is to simply shut off some of the alarms. Although this is easy in terms of actual effort, it can be mentally difficult for a DBA to turn off alarms out of fear they will miss a critical issue notification.
To overcome that fear, spend some time analyzing alerts to determine those that are chronically false alarms, those that don’t need immediate attention, and those that can be handled with a bit of automation. Once you have identified these alerts, you can feel confident in turning them off and tracking them via logs.
Now that you have the noise under control, it’s important to review all of your SQL Server alarms to ensure they are set up correctly. Properly set up alarms are crucial not only to the performance of your server but also to your sanity.
To get started setting up efficient alerts, especially if you inherited the DBA job, first look at the existing alarms with fresh eyes. Do you see anything unusual? If you’ve been been the DBA for a while, you may want to start by reviewing your logs to see if there are trends indicating some maintenance is in order.
Once you have a good feel for the state of your existing alerts, there are some best practices you can implement that will ensure you are receiving only the most critical alerts and those that need immediate human intervention.
Here are four tips for setting up effective SQL Server alerts.
Set Specific Rules for Alarms
This is a great way to avoid unnecessary alarm noise. You don’t need to know all the things all the time. Setting specific rules, such as “exclude instance X from the alarm” or “only send an alert if this alarm is triggered Y times,” will help you zero in on the most important issues and save you from alarm fatigue.
Use Smart Alarms That Not Only Send Alerts but Also Start the Diagnostic Process
There are some SQL Server performance monitoring tools on the market that don’t just tell you that you have a problem, they also give you ideas about how to fix it. If you isolate the potential root cause of an alarm right off the bat, your resolution time will likely be significantly shorter.
Prioritize Alarms so You Tackle the Most Important Problems First
The faster you identify a performance-killing issue, the better. Be sure your SQL Server monitoring tool lets you customize and prioritize alarms so you can easily see the really big problems fast. You decide the threshold, when to sound an alert, what to ignore, and what action to take when an alert is triggered.
Take Advantage of Mobile Alerts
Mobile performance monitoring has done amazing things for DBAs’ work-life balance. Alerts can now be assessed and sometimes even cleared from any device and any location, so DBAs are no longer at the mercy of the 3 a.m. phone call. Some mobile monitoring apps even offer heatmaps that provide at-a-glance data on the severity of the issue triggering the alert.
Now that you have some ideas for setting up your alerts so you spend less time reacting to issues that don’t need immediate triage, find out how Spotlight Cloud makes SQL Server performance monitoring even easier.
Spotlight Cloud’s cloud-powered SQL Server performance monitoring solution has a wealth of features designed to simplify performance monitoring. Some specific features that will help you set up efficient alerting include:
More Than 100 Defined Alarms
Spotlight Cloud has more than 100 defined alarms that alert the user when they are triggered. Because having 100+ alarms sending you alerts at all hours is the opposite of efficient, Spotlight Cloud’s alarms are customizable so you only receive alerts for alarms that you specify.
Health Check Screen
Monitoring server health is critical to pinpointing existing and potential issues that can affect performance. Spotlight Cloud’s Health Check screen provides data on key system performance metrics, helps isolate the root cause of issues, and lets you drill down to get details and plan a course of action.
Noise Reduction
SQL Server performance monitoring can get noisy. Spotlight Cloud comes with pre-configured alarms designed to help you prioritize so you aren’t overwhelmed by alerts. These alarms are customizable to meet your organization’s specific needs, so you won’t be bombarded with notifications you don’t need.