When the Agile Manifesto was written over 15 years ago, it focused on four main values and produced the greatest paradigm shift the software development industry has even known.
- Individuals and interactions over processes and tools.
- Working software over comprehensive documentation.
- Customer collaboration over contract negotiation.
- Responding to change over following a plan.
But What About the Bugs?
In the utopian agile concept (and in Utopia in general), there would not be bugs.
In the reality of development, however, things come up. Errors are discovered. Mistakes happen. Agile development teams need a way to handle bugs that doesn’t interfere with their agile process–moreover, it needs to empower their agile process.
Avoiding Bugs: Individuals and Interactions
One of the best ways we deal with bugs is empowering all of our individuals developers to avoid bugs in the first place. We continuously deploy and use test-driven development, but one of our greatest assets is one of individuals and interactions: A rigorous code review process.
When devs interact in this way, it ensures really tight code with minimal defects (often, none at all).
Our developers submit code from a branch, and team members can pull and test changes, view change sets and affected files, submit a new version, vote on the request, and ultimately merge or reject the request.
Having multiple individuals interact in this way is one way to interpret the agile manifesto to avoid bugs.
Resolving Bugs Quickly: Working Software
We don’t have a documented policy about dealing with bugs–we just do it. And we do it quickly, especially when it comes to blocker bugs, to make sure our software works. We typically deal with bug fixes as they are discovered and include them in the current sprint whenever we have the capacity. When we are not able to solve the issue right away, we can easily roll back to a previously deployed working version within eight minutes of discovering a bug.
While there is no single approach (documented or not) that works for all teams on all projects, our preference is always to fix sooner rather than later, maintain lean inventory, and always ensure the product is working.
Customer Success and Test Cases: Customer Collaboration
We are fortunate in that many of our customers are developers themselves. In the (uncommon) case that a customer of ours discovers a bug and reports it to our Customer Success team, we work to collaborate with them on understanding the issue and fixing it quickly. We then write test cases. We will never fix a bug without writing a test case for automatically catching such a bug in the future.
Post-Deploy Monitoring: Responding to Change
On our teams, we have found that users find new ways to customize our software to their workflows. We may design a feature for a certain use case but only when it is used in a completely different way do the bugs surface. These ongoing changes can only be responded to with post-deploy monitoring.
This is our way of responding to the changing needs of our customers and the changing uses of our products.