Reliability starts with the application
The development-first approach to resilience
Start testing for reliability early and often by synthesizing reliability tests from the tests your developers are already writing
Developers write standard integration and functional tests
Filibuster automatically synthesizes test variations where faults are injected in application
Developers update tests to assert the correct behavior under fault
Built on peer-reviewed software engineering research
Filibuster is based on published research that demonstrates that many incidents are caused or exacerbated by the incorrect error handling
Research published at the ACM Symposium on Cloud Computing in 2021
Evaluated with real incidents taken from industrial, microservice applications
Developed in the Software and Societal Systems department at Carnegie Mellon University
Build consumer trust from a culture of reliability
Ensure that when faults inevitably occur, your application provides the correct user experience under those circumstances
Verify application error handling works correctly
Verify you gracefully handle non-fatal errors
Ensure correct messaging when fatal errors occur