Posted September 03, 2015
How Smart Are Your Performance Alerts?
Performance alerts can be a double edged sword. The ‘who, what, when and how’ of performance alerting decides the overall effectiveness and ability to avoid potential disaster. For example, if you are a web master, you don’t want to receive alerts at 3:00am when you know the website is down for scheduled maintenance. Or if you are the Chief Information Officer, you don’t want your inbox flooded with redundant alerts.
As web and mobile applications go all-out to deliver best user experience, the environment that supports them becomes more complex than ever. Now instead of one person or one team responsible for your ‘online-presence’, you have different teams, possibly working from different geographical locations, perhaps even employed by different companies, managing your web and mobile applications. And then there are the C-suite personnel and decision makers who need to be alerted when there are issues of serious business impact, but may not necessarily appreciate dozens of emails, texts or phone calls alerting him or her on the same thing.
What we need here is an intelligent alerting system that is reliable, responsive and fast. One that helps you to:
- Detect and avoid issues before your users encounter them
- Get alerted on availability as well as performance issues
- Filter signal from noise by avoiding false positives
- Receive alerts that need YOUR attention
- Get/Stop alerts via the means of your choice
Who should be alerted?
This is kind of a no brainer. Those who are responsible to put down the fire should get alerted on it. However not all service providers let you select recipient(s) based on nature of the issue. An intelligent alerting system solves this problem; it lets you configure alerts by error types. So the right people get notified at the right time, resulting in faster resolution of the issue.
What you should be alerted on?
Availability and performance:
Your web or mobile applications may be available in your eyes, but if they are slow or don’t let users do what they want to, they are as good as unavailable in your users’ eyes. Intelligent alerting systems not only alert you on availability issues but also on performance issues.
Static vs dynamic:
Many vendors allow you to configure static i.e. fixed performance thresholds and alert you when the threshold is reached. This is great! But static thresholds don’t alert you about unique unusual behaviors in your environment that may need your attention. What you need here is dynamic threshold, which takes into account the behavior over past few days and sets accepted upper level and lower level threshold. This way, you get alerted when there is any unusual behavior in your environment, even though it is within your accepted threshold. An ideal alerting system allows you to customize the performance thresholds and gives you a choice to configure static and dynamic alerts.
If you have interactive web or mobile applications, some steps of the user transaction are more important than others, such as submitting a form or checking out with credit card. If those steps fail due to internal or third party error, you should be immediately notified.
An intelligent alerting system lets you configure alerts on a particularly important step, giving you more insight into user experience and possibly identify reasons of performance degradation.
When you should be alerted?
Alerting is only useful if it gives you enough bandwidth to understand the issue, gather experts and resolve the issue as quickly as possible. By definition, alerts should be fast! However an intelligent alerting system performs the check again to avoid false positives.
Mind the blackouts
However there could be some set time frames where you don’t want to receive alerts. This seemingly unlikely scenario is in fact very common with the global nature of operations. Your operations teams could be spread across different locations and take turns to monitor your applications, members of the team might go on vacation or business trips, or you may have a set on call rotation for after-hours issues. And then there are scheduled downtimes, such as maintenance or planned network outages, when you don’t want to receive alerts.
An intelligent alerting system lets you to route the alerts based on recipients’’ work-hours and set blackout periods for planned activities.
How you should get/stop alerts?
Alerts are only useful if you get them via the mode of communication most appropriate to you. Some of us prefer email messages, while others may prefer phone calls or texts. Some use sophisticated services like PagerDuty that helps you with alert aggregation and dispatching. An intelligent alerting system lets you to select the alerting means of your choice.
On the other hand, you don’t want to be alerted on the same issue if you have already started working on the resolution. An intelligent alerting system lets you turn off the alerts remotely, again, by your choice of mode of communication.
We would like to hear how you define the ‘who, what, when and how’ of alerting. Let’s start expecting more from the alerts. Sign up for a free AlertSite trial today and see how we do alerting intelligently.