Balancing Reliability and Feature Delivery Using Error Budgets
Esther, the product owner, analyzes user data and discovers a decrease in users. After researching user experiences, Esther discovers that users are moving to competing products with more features.
In a retrospective, Bob notes that deployments have many manual steps and verifications to prevent errors. The developers think that complex deployments affect their velocity.
Esther works with Alice and learns that the product averages four minutes of downtime per month, or has 99.99% availability. Esther's user experience research reveals that users can tolerate up to 45 minutes of downtime per month, or 99.9% availability.
Esther, Bob, and Alice establish an error budget of 40 minutes of downtime per month.
Alice and Bob work on the deployment process. Changes to the deployment process cause some downtime, but the downtime does not consume all of the error budget.
Improvements to the deployment process help the development team spend more time developing features. The organization learns that embracing acceptable downtime helps increase velocity to match competitors.
Reflect about the preceding history.
Post your opinion about what would happen if a product owner proposed this change in an organization you know. Review other students' posts and discuss.