Flipkart: Scaling Aerospike at Flipkart with Kubernetes
summary
Flipkart centralizes 170 Aerospike clusters under a DBaaS team, migrates VM deployments to Kubernetes, and runs Aerospike containers directly on bare metal for bare‑metal performance while using non‑root containers and multi‑tenant namespaces. The Aerospike operator handles seamless horizontal and vertical scaling including PVC scaling, CI/CD drives cluster lifecycle as code, and a monitoring/backup ecosystem provides fragmentation repairs, billing and domain‑aware alerts; the setup reports 100% availability with no operator-caused incidents.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 0:49Business Use Cases
- 1:49Problems with VMbased deployments
- 2:55DBAS
- 4:32Enterprise License
- 5:55Deployment Strategy
- 6:56Policies and Guard Rails
- 8:56Manageability
- 10:23Kubernetes Operator
- 13:07Deployment Life Cycle
- 13:49Ecosystem
- 16:52APJ Architecture
- 18:56Monitoring Engine Features
- 22:45Cluster Visibility
- 23:11Hardware Visibility
- 23:48Future Enhancements
- 24:40Sustainability
related talks
50mRunning Databases Efficiently at Scale - Case Study of Flipka...
27mPods on Steroids - A Case Study of Migrating Flipkart.com to K8s on Bare Metal
06mSnapdeal and Aerospike
49mHow Flipkart runs HBase across zones and clouds seamlessly
21mCase Study: FOSS (Apache Hbase) at Flipkart
35mAerospike on IBM Cloud: Real-time Workloads Demand Real Performance
16mBuilding a K8s Operator for a Distributed Database
21mBuilding a K8s Operator for a Distributed Database