DoorDash Diaries: Mastering Microservices at Mega Scale
summary
DoorDash runs thousands of microservices and a >1,000-node call graph that handles ~574 million orders a quarter (peak ~590 orders/sec → ~35k RPCs/sec, overall >500k RPCs/sec). Stack is Kotlin/Node/Python with Kafka, Elasticsearch, Redis and OpenTelemetry; service mesh plus tenant-aware Redis shims enable multi-tenant routing and hard‑swapping individual nodes so teams can run testing-in-prod. Platform enforces zero-touch deployments, SLO-driven automated rollbacks and standardized libraries to limit tech sprawl and preserve developer velocity.
AI-written summary · about
chapters (5)
- 0:00Introduction
- 0:39What is DoorDash
- 3:55Breaking Down the Monolith
- 8:57Testing in Fraud
- 13:12Platform Standardization
related talks
23mMicroservices – Diptanu Choudhury
38mHighly Scalable Consumer Applications using Cloud Native Open Source Tools
22mBuilding + operating a service mesh at a mid-size company
23mDevOps at Uber
28mMicroservices – Boyan Dimitrov
26mFighting the Good Fight at the Hot Gates of Micro-Services
31mMonolith Decomposition
46mMETRO's Journey into Microservices: Experience and Lessons learned along the way