Load Balancing in gRPC
summary
Explains gRPC client-side load balancing with the actual LB policy API, the plug-in builder/balancer/clientConn/picker interfaces, and how service config service config names LB policy drives runtime selection; shows subchannel lifecycle and a common state-aggregation rule (ready > connecting > idle > transient failure); defines picker semantics and fast-path return types picker returns complete queue fail drop and notes the Java/Go APIs are experimental while implementations remain stable.
AI-written summary · about
chapters (7)
- 0:00Load balancing in gRPC
- 1:24Server vs client-side
- 3:10Load balancing policies
- 6:43gRPC client load balancing
- 10:12Subchannels and states
- 13:01Load balancing API
- 17:58Q&A
related talks
25mLoad Balancing in gRPC
01h 01mgRPC client-side load balancing — why, what, how and WTF?
33mCloud Native Load Balancing From Scratch
29mgRPC load balancing and Service Mesh
35mgRPC Load Balancing on Kubernetes - Jan Tattermusch, Google
18mDeveloping a Custom Load Balancer to serve traffic at scale
33mCloud Native Load Balancing From Scratch
23mDemystifying Load Balancing in Istio
