Tips and metrics for application monitoring

Photo by Geralt from

Want to make sure your daily backups are running? Take a look at PushMon, the easiest way to monitor your cron and scheduled tasks.


There is a digital revolution happening in all sectors of society, driven by consumers who are increasingly eager for innovation. In a market driven by the maxim “time is money”, it is practically impossible not to fail for lack of good performance. So how do you stay competitive in the face of these rapid transformations, demands and growing expectations of users? Before launching or presenting any product or service, it is imperative that all processes are structured and fit. For this, a good idea is to use DevOps practices to support a continuous delivery model.

A survey by Puppet Enterprise points out that high-performance DevOps users are more agile, deployments 30 times more frequent and 8,000 times faster than the traditional system. They are even more reliable, with 12 times faster recovery if there is a problem. In addition to this model, it is important to follow some relevant metrics and tips on how to launch an application that provides a responsive, responsive, and most importantly, digital experience that does not fail at the time of its use.

Do not force without a plan

If your applications are not ready, are slow or have crashes, do not go forward. The company’s best campaign, which has been invested the most time and money, can turn out to be the worst of them and damage the company’s image if there is a flaw in its presentation. To give you an idea, during a major sporting event, the organizer’s mobile-developed website featured a favicon (small icons that stand next to a browser’s address bar and serve to quickly identify a website) with size of 370 kb, when the normal is to be between 512 bytes and 2048 bytes. Obviously the action was seriously compromised and the case could have been avoided if there was a basic optimization of web performance and testing throughout the development pipeline.

Do not assume that you know the environment

In an application, however small, there are many potential points of failure. The multiple devices, technologies, channels, and methodologies exponentially magnify the possibilities of something going wrong. Therefore, it is important not to assume full knowledge of the environment without the actual perception of a user.

Be careful when reuse components

Developers are always reusing existing components, but this does not always work for the company. It is advisable to closely follow all the processes.


Key performance metrics include number and size of features, page size, number of functional errors, third-party calls, number of SQL executions, and number of SQL statements. Other highlights include time spent on APIs, API calls, number of domains, total size, number of items per page, and AJAX per page.

Ideally, you should control these metrics manually throughout your application development pipeline. Once there is control over what you need to know, it’s time to start looking at how to simplify performance monitoring. This is the goal of continuous delivery: automate your procedure with quality portals based on metrics at each step.