If you translate the concept of a car dashboard into the digital environment, monitoring tools, with all their graphs and pointers, indicate how an (online) application or software programme performs. This provides engineers with answers to the questions regarding which type of defects are present and where maintenance/action is needed. However, this approach no longer seems to be appropriate for current practice. After all, a technical 'alert' does not necessarily have to be a red flag for business operations. These two worlds therefore need to merge. Fortunately, an increasing number of tools are available in response to that need.
Monitoring tools such as Datadog, which we often use at SQLI, are increasingly bringing technology and business value together. For example, a dashboard shows how stable an application is, whether processes are slow and whether a database is filling up. But the dashboard also looks at the number of users currently active, the number of orders placed and the error messages that occur for end users. This takes into account, among other things, customer journeys (simulations of customer behaviour) and how parts of these can be correlated with the technical signals.
In the past, this meant that everything worked in the back-end, but a problem at the end-user side resulted in a drop in the number of orders. Conversely, a technical problem can be given top priority, but if the inflow of customers does not start until 10 in the morning, it is not absolutely necessary to have solved the problem in the middle of the night. By combining these information flows and viewing reality in a consistent manner, it is possible to include the business impact in the assessment of a technical problem.
The big picture
This not only leads to better decisions that are right for the organisation as a whole, but the engineers are also better able to do their work. If, in the past, they had to examine four applications at the same time to find information, a single interface now provides a complete picture. Equally as important, the alerts that are configured take into account more variables, thereby reducing the number of unnecessary alarms. As a result, there is less chance of 'alert fatigue', i.e. that the engineers receive so many signals that they become desensitised to them.
The merging of technology and business leads to more pro-activity with a focus on a shared interest – the success of the company. The development links up perfectly with that of DevOps (a combination of Development and Operations). One of the aims of this collaboration model, which brings together developers, product management and operations, is to operate in a more agile manner. The monitoring tools described support this. With each new release, the effect of the code is distinctly clear. If it does not appear to have been written well and an important metric, such as loading time, is adversely affected, the DevOps team is aware of this in the test environment. All in all, it is easier to develop and innovate software pro-actively and more effectively.
The business through the eyes of the client
Monitoring tools are costly. This makes it all the more important to stay on top of things and to use information effectively. An important key to success lies in the proper configuration of the alerts and assigning the right priorities. A simple example: if an e-commerce company works with a late cut-off time of 11pm ('Order today for next-day delivery'), every error in the order flow before that time has a direct impact on the business. Such an error must therefore be given the highest priority. However, a problem after that time is less urgent.
In order to make a good analysis, it is wise to involve your client in the monitoring process. For example, we always choose to hold extensive interviews that give clients the opportunity to familiarise us with their business. What issues have the greatest impact? Which delicate processes are familiar to the IT staff? And where does the support department often identify failures? This kind of approach makes it possible to determine which indicators from business operations require attention and which technical aspects relate to them.
A monitoring strategy that takes into account both worlds not only prevents business and other problems early on, but enables innovation. Being pro-active and bringing together both worlds creates opportunities to experiment, improve features and innovate.
This article was originally published at EMERCE in Dutch.