Open Source Summit + ELC North America 2020

eBay runs dozens of Kubernetes clusters across global data centers in different regions. Tens of thousands of nodes support eBay core services such as search and big data.

With more applications getting onboarded to eBay cloud platform, the infrastructure reliability is becoming critical for both platform and applications. We keep thinking ways to better manage the lifecycle of eBay infrastructure.

Since systems and hardware can fail, we are providing a unified and highly automated infrastructure remediation system in eBay, which is fast and reliable, with customizations for different customers and workloads, to remediate various failures, manage breakfix actions without affecting running workloads, and ensure resources coming back, so as to manage the complete infrastructure and application lifecycle for availability, reliability, and durability.

In this talk, we hope to share how we modeling eBay cloud infrastructure, and building the unified cloud infrastructure remediation workflow.