Kubernetes SIG Instrumentation - Deep Dive
Talkat KubeCon + CloudNativeCon North America 2019 · on Wednesday, 2019-11-20· at 15:20· Room 6D - San Diego Convention Center, Upper Level·
summary
Kubernetes adds a concrete metrics lifecycle: a metric stability framework that tags metrics with stability and deprecation, enforces registration through wrappers and static analysis, and auto-hides deprecated metrics for one release so teams get one release to migrate. It also prototypes distributed tracing using OpenTelemetry/Zipkin/Jaeger by propagating trace context in object annotations, producing one trace per object that links API server, controllers, kubelet and CRI calls for end-to-end debugging; demos show it working on pods and deployments.
AI-written summary · about
chapters (23)
- 0:00Intro
- 0:15What we are going to cover
- 1:04SIG Instrumentation
- 2:56Bad Metrics
- 4:25Fixing Existing Metrics
- 5:32Metrics as an API
- 6:38Alternatives considered
- 7:39Goals
- 8:25Quasi-Versioning Strategy
- 9:16Prometheus Metric Lifecycle
- 10:41Hijacking Metric Definition
- 11:07Hijacking Metric Instantiation
- 11:23Hijacking Metric Registry
- 11:49Stability Axes (axises)
- 13:19Enforcing Stability
- 16:18Story Time
- 17:32Tracing in Kubernetes
- 21:44Span Context Propagation
- 24:56Tracing a Pod Creation
- 25:53Tracing a Deployment Creation
- 29:50Tracing, Generalized
- 32:22This is a Work-In-Progress
- 35:39Open Telemetry
related talks
28mIntro and Deep Dive: Kubernetes SIG Instrumentation
26mIntro and Deep Dive: Kubernetes SIG Instrumentation
30mIntro: Kubernetes (Instrumentation) SIG
24mSIG Instrumentation Introduction and Deep Dive
40mSIG Instrumentation Introduction And Deep Dive
35mSIG Instrumentation Introduction and Deep Dive
36mSIG Instrumentation: Intro, Deep Dive and Recent...
28mSIG Instrumentation Introduction and Deep Dive