How to Find Bottlenecks in the Agile and DevOps Delivery Cycle
by Tanya Kravtsov
The iterative, collaborative models of DevOps and agile development changed the way we deliver software. But adopting these models requires more than just continuous integration, continuous testing, and continuous delivery — it requires a complete mind shift.
In order to achieve incremental software development and continuous feedback, you need to eliminate the tasks that create waste or bottlenecks, which hinder the flow of development. A chain is no stronger than its weakest link, and identifying these “weak links” is a critical step toward achieving agility and increasing efficiency.
The first step is gaining a full understanding of the system, including all the players and the handoffs between them. Mapping the delivery pipeline in the form of a flow chart should quickly highlight the points of improvement.
Talking to people from different groups within the organization who play a role in the system also can go a long way toward understanding where the bottlenecks are. Interviewing members of the development, quality assurance, product, IT operations, marketing, and senior leadership teams will identify the common themes and present ideas for improvement.
Brainstorming techniques such as mind maps, storyboarding, and innovation games such as Speed Boat can help facilitate these discussions. Retrospectives, an essential part of agile Scrum practices, are another great tool that encourages teams to provide their feedback on what works and what doesn’t
The most common bottlenecks in the software delivery lifecycle often turn out to be environments, testing, and communication. (How many times have we heard “But it works on my machine”?)
To avoid such issues, it’s important to automate the environment setup process by identifying all the steps necessary to set up a machine and scripting them using infrastructure as code and configuration automation tools, in conjunction with service virtualization. It’s also essential to either give developers and testers the ability to spin up preconfigured environments on demand or to have dedicated, stable environments with consistent monitoring and alerts set up in case something doesn’t work as expected.
Testing is often considered a bottleneck due to its manual nature, and this is compounded when testing starts at the end of the development cycle or sprint, which often impacts sprint commitments and delivery dates. The best way to combat this is by incorporating testing into the development process and automating everything that can and should be automated, such as unit, regression, and performance testing.
There are great static analysis tools that can be integrated into the developers’ pipeline to validate the quality, security, and unit test coverage of the code before it even gets to the testers. It’s also important not to overlook tedious and unnecessarily manual processes, like test data gathering and test results analysis.
Even when all the technical bottlenecks are addressed, miscommunication and lack of transparency can completely break the flow and become huge blockers in an otherwise continuous and streamlined process. Defining ownership, understanding your audience, creating clear and simple messaging, and managing expectations are critical tools in breaking the communication barrier and moving toward a collaborative, agile organization.
Originally published at www.techwell.com.