Navigating the Transition: SRE Challenges and Highlights
summary
Adidas migrates its e-commerce stack from a monolith to microservices running on Kubernetes + AWS, with Kafka, Jenkins, OpenTelemetry, OpenSearch and an APM. Throughput jumps from ~3–4K orders/min to 40,000+ orders/min, but Black Friday exposes a coupon DB bottleneck that causes a 90‑minute outage; an SDK version mismatch later adds a 35‑minute outage. CI/CD lives in per‑service Jenkins pipelines with duplicated Terraform and Helm code, feature flags and canary deploys reduce blast radius, and incident management uses an Incident Commander, revenue‑impact KPIs, blameless RCAs and tracked remediation.
AI-written summary · about
chapters (8)
- 0:00Introduction to SRE transition
- 1:09Reasons for shifting to microservices
- 3:37What we moved to
- 6:53CI/CD automation setup
- 9:43Key SRE challenges faced
- 19:37Building resilience after failures
- 23:10Incident management process
- 27:44Continuous journey in SRE
related talks
41mBuilding and running applications at scale in
37mDelivering agility at McDonald's with microservices transformation
51mCompetitive retail architecture with microservices
38mFrom Monolith to Microservices at Zalando
51mMicroservices and Microfrontends: Scaling, Learning, Growing
29mMoving from A and B to ~150 microservices, the journey and learnings
01h 07mMicroservices from the Trenches. A Guide through the Architecture, Design, Code
59mJava microservices from Netflix OSS to Kubernetes