Infrastructure Monitoring: Differentiating Actionable Alerts from Noisy Ones

Estimated reading time: 3 mins

As of March 2018, Google Play store hosted 2.6 million apps, according to Statista. From the perspective of the competitive business world, this means that it is now even more competitive to launch a business app and have it used consistently by your target audience. Immediately after users notice a hiccup in the functioning of your mobile application, most of them will lose trust in using it, while others will even delete it from their phones.

In the quest to remain relevant in this ever-competitive sphere, infrastructure monitoring can prove to be valuable in identifying and dealing with app loopholes as soon as they occur. Given that there are a lot of metrics that could trigger app monitoring alerts, this situation results in noise which makes it tough to differentiate the valuable and sensitive alerts from those that need little attention. Failure to make out the difference often results in false positives where you might alert software developers of an issue that is non-existent.

Here is how to optimize your alert systems to differentiate between sensitive and non-sensitive alerts:

Have Different Levels of Alert Urgencies

Different alerts from your infrastructure monitoring tool will call for different levels of attention –  hence not every alert has to be immediately reported to the engineers. For instance, an alert that arises from network congestion is only but transient. The problem is bound to go away by itself. As such, you should prioritize alerts with regard to the kind of response urgency that they require. Here is how to do so:

1.    Place Low Severity Alerts as Records

Some alerts will only affect the service level of the application and infrastructure. They will barely affect how the app functions. For instance, user service queries might start being served at a slower rate than usual, but the change might be too minimal to cause any downtime. In such cases the condition should be stored in the records for future reference and contextual investigation. In case the issue gets to the point that it leads to a decline in application performance, then the stored records will provide valuable information during the investigation stage.

2.    Use Notifications for Moderate Severity

While some alerts will require the intervention of IT professionals, they might not necessarily be emergencies. For instance, alerts that storage is running out within a week will need to be addressed but not with immediate effect. Instead of making IT professionals drop their duties to attend to such alerts, they should instead be sent through notifications such as emails to alert them of any issue. These notifications will be visible to them but not disrupting, especially in situations where the IT officer might be asleep.

3.    Use Alarm Alerts For High Severity Issues.

Some issues require to be addressed as soon as they arise. For instance, unfavorable app latency should be looked into regardless of the hour in which it is discovered. To avoid the potential loss that ignoring such an issue can spawn, it is vital to alert the responsible individuals as early as you can. This can best be done by sounding an alarm in the phones of such individuals to command their attention wherever they may be.

Early Warning Signs Should Not Be Ignored

There are cases where an alert might signify an early warning sign of a subtle problem that could easily grow into a major issue. As a result, such warning signs should never be ignored. For instance, a sign that you are running out of server memory space should be reported early. Though this is not an issue to wake the IT professionals in the middle of the night, early reporting will give them enough lead time to deal with this issue.


App problems should be dealt with in order of priority, and false alarms make it hard to achieve this. Working with a context-aware application performance management system will help make alerts effective enough. Set up your alerts as outlined above to optimize the insights you receive from your APM tools.

Check out these similar posts:

Leave a Comment

Please note: if you are making a comment to contact me about advertising and placements, read the Advertisers page for instructions. I will not reply to comments about this subject.

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Scroll to Top
How Am I Doing?

Did this discussion solve your problem?

Then please share this post or leave a comment.