When adopting DevOps, many organizations focus on increasing the speed of software delivery and deployment without ensuring that they have adequate breaks in place. As a result, instead of reaching the promised benefits of DevOps, these organizations often end up “hitting the wall.”
๐ช๐ต๐ฎ๐ ๐ฑ๐ผ๐ฒ๐ ๐ฎ ๐ฟ๐ฎ๐ฐ๐ฒ ๐ฐ๐ฎ๐ฟ ๐น๐ผ๐ผ๐ธ ๐น๐ถ๐ธ๐ฒ?
When this question is asked, many think of sleek car designs, powerful engines with immense horsepower, and smooth-shifting gearboxes. Since race cars imply speed and acceleration, it is perhaps natural that such features come to mind directly
However, race cars also have high-performing braking systems, often more powerful and superior to those in standard consumer vehicles. The braking system allows the car to maneuver around corners securely, decelerate when turning, and stop safely when needed. Just imagine the consequence of the brakes not being appropriately adapted. Would anyone dare to hit the gas pedal if they had all other race car features?
Therefore, contrary to many beliefs, ๐๐ต๐ฒ ๐บ๐ฎ๐ถ๐ป ๐ฏ๐ฒ๐ป๐ฒ๐ณ๐ถ๐ ๐ผ๐ณ ๐๐ต๐ฒ ๐ฏ๐ฟ๐ฎ๐ธ๐ถ๐ป๐ด ๐๐๐๐๐ฒ๐บ ๐ถ๐ ๐๐ต๐ฎ๐ ๐ถ๐ ๐ฎ๐น๐น๐ผ๐๐ ๐๐ผ๐ ๐๐ผ ๐ด๐ผ ๐ณ๐ฎ๐๐๐ฒ๐ฟ! In addition, ๐๐ต๐ฒ ๐ณ๐ฎ๐๐๐ฒ๐ฟ ๐ฎ ๐ฐ๐ฎ๐ฟ ๐ถ๐, ๐๐ต๐ฒ ๐บ๐ผ๐ฟ๐ฒ ๐ฎ๐ฑ๐ฒ๐พ๐๐ฎ๐๐ฒ ๐ฏ๐ฟ๐ฎ๐ธ๐ฒ๐ ๐ถ๐ ๐ป๐ฒ๐ฒ๐ฑ๐.
The exact relation between carsโ brakes and speed can be applied to ๐๐ฒ๐๐ข๐ฝ๐. For many software development organizations, DevOps is about โspeed and acceleration,โ leading to quick market time, short lead times, and frequent software deliveries. Thus, these organizations see DevOps practices such as continous integration and delivery (CI/CD) as the โgas pedalโ for faster delivery and deployment. While this is true, many organizations often underestimate – or even neglect – the need for an adequate DevOps braking system that allows for increased acceleration and speed to be reached.
๐ช๐ต๐ฎ๐ ๐บ๐ฎ๐ธ๐ฒ๐ ๐ฎ ๐๐ฒ๐๐ข๐ฝ๐ ๐ฏ๐ฟ๐ฎ๐ธ๐ถ๐ป๐ด ๐๐๐๐๐ฒ๐บ?
The DevOps braking system is made of ๐๐ต๐ฒ ๐ณ๐ฒ๐ฒ๐ฑ๐ฏ๐ฎ๐ฐ๐ธ ๐น๐ผ๐ผ๐ฝ. The evolution of the feedback loop needs to be tightly coupled with the speed of software deliveries and deployment. To illustrate this, letโs assume that a software-as-a-service organization deploys a new software release once a month to its production system. While the organization monitors the production environment continuously, the average time it takes to resolve a software bug is 3 weeks (including time to detect, respond, and repair).
Thus, if the organization considers accelerating the delivery and deployment cycle to one month, it must first ensure that the average bug resolution time is reduced to 2 weeks or less before attempting the acceleration.
๐ง๐ต๐ฒ๐ฟ๐ฒ๐ณ๐ผ๐ฟ๐ฒ, ๐ถ๐ณ ๐๐ผ๐ ๐๐ฎ๐ป๐ ๐๐ผ ๐ด๐ผ ๐ณ๐ฎ๐๐๐ฒ๐ฟ, ๐บ๐ฎ๐ธ๐ฒ ๐๐๐ฟ๐ฒ ๐๐ผ ๐ต๐ฎ๐๐ฒ ๐ฎ๐ป ๐ฎ๐ฑ๐ฒ๐พ๐๐ฎ๐๐ฒ ๐ฏ๐ฟ๐ฎ๐ธ๐ถ๐ป๐ด ๐๐๐๐๐ฒ๐บ ๐ณ๐ถ๐ฟ๐๐.
What happens in many organizations is that they often attempt to increase the speed of deliveries without ensuring there is a proper feedback loop in place. As a result, after a short period of acceleration, the organization often realizes that things have gone out of control, which can be realized with stressed developers juggling multiple tasks simultaneously, unsatisfied customers, and more complaints.