Filling the Gaps in Kubernetes Test Coverage - Hippie Hacker, ii.coop
summary
Kubernetes conformance coverage is measured by an audit-log pipeline called API snoop that parses audit events, stores them in Postgres and a GraphQL layer, and correlates them with TestGrid/Prow runs; current stable APIs show roughly 30% coverage with a ~70% gap, about 1,400 endpoints total and ~400 unstable operations excluded. The demo deploys dynamic auditing, runs e2e tests (Go/Ginkgo, plus Python/JS examples), and visibly increases coverage by a couple of points.
AI-written summary · about
chapters (12)
- 0:00Welcome
- 1:48Executable ideas and orchestration
- 3:53New standards for workloads
- 5:01Kubernetes API surface
- 6:49API coverage history
- 8:43API test coverage evolution
- 11:20Key contributors and policy
- 12:04Get involved in test coverage
- 12:27Demo setup and API Snoop
- 17:16Running tests and coverage impact
- 18:46Improve Kubernetes test coverage
- 19:38Q&A
related talks
21mFilling the Gaps in Kubernetes Test Coverage
34mKubernetes Conformance Deep Dive
23mContributing to Kubernetes Conformance Coverage
16mContributing to Kubernetes Conformance Coverage
20mConformance Testing the Kubernetes API: Tooling that Makes the Difference
27mAdventures in Conformance
35mK8s Conformance & SIG Architecture Conformance Subgroup
01h 27mIntro + Deep Dive: Kubernetes Conformance WG
