Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Learn how to manage flaky test cases in the SONiC network operating system through this 35-minute conference talk from the Open Compute Project. Discover the challenges of handling over 1,000 test cases that run nightly across various hardware SKUs and platforms, and explore how Microsoft's internal Nightly Hawk system automatically surfaces issues from test runs. Understand the critical distinction between consistent failures that are reproducible and easier to debug versus flaky failures that are intermittent and often overlooked, which can mask real issues or signal test instability. Explore a comprehensive system designed to manage flaky tests by collecting and tracking them across multiple runs, categorizing different failure types through summary analysis, surfacing issues without duplication, and providing clear indicators to help prioritize repairs in development workflows. Gain insights into streamlining failure detection processes and improving the overall reliability of network operating system testing infrastructure.
