Confident Canary Deployment to Production With Istio
summary
A Mexico fintech runs Istio across a cluster that grows from 5 to 80 services in 18 months and uses header-based routing to direct internal users to gray deployments while live users hit green, avoids twin stacks by swapping traffic with 30/70 rollouts (also 25/75) and gates releases on error rates and business metrics. Observability uses Kiali and Jaeger plus Prometheus dashboards; Kafka topics are duplicated per environment; complexity grows with configs, so Helm centralizes manifests and it works.
AI-written summary · about
chapters (13)
- 0:00Intro
- 0:16Raju Dawadi
- 0:28About oyster.io
- 0:58Istio at our Heart
- 1:43Deployed" VS "Released
- 2:27In Early Days...
- 2:58Felt Like Monolith
- 3:26Grey Service Plug-in
- 4:40Istio Rule
- 5:19Canary After Internal Test
- 6:06Challenges We Faced
- 7:38Extra Confidence
- 9:30Too Many Yaml
related talks
13mCanary in the Internet Mine
18mMicroservices Out In the Wild
29mK8s Complexity Tamed: One Interface To Rule Them All With CNCF Incubat...
45mHow Kubernetes Was the Secret Sauce in Our Globally Distributed Team’s Transition to Microservices
45mGrowing at the Edge: Doubling Traffic While Changing the API Gateway
39mEmbracing Cloud Native at a Thriving, Established Company
33mKubernetes at Cruise: Two Years of Multitenancy
31mUnity in Diversity: Architecting “Shared-Fir...