Core concepts
MinIO Sidekick provides client-side load balancing with automatic health monitoring and multi-site failover capabilities.
Traditional load balancing architectures introduce an additional network hop between clients and servers, creating potential bottlenecks and single points of failure. MinIO Sidekick eliminates this intermediate layer by running as a lightweight sidecar process directly alongside each client application.
Sidekick maintains awareness of backend server health through active monitoring and intelligent request routing. Health check probes run in the background, evaluating each backend’s readiness to serve traffic.
When servers fail or become degraded, Sidekick stops routing requests to those endpoints. This failover detection ensures that clients experience minimal disruption during backend failures, as each sidecar independently adapts to changing server availability.
| Topic | Description |
|---|---|
| Architecture | Sidecar model, request routing, and component structure |
| Multi-site failover | Automatic failover between sites and read-only health checks |
| Load balancing | Least connections and random distribution algorithms |
| Health checks | HTTP health monitoring, status transitions, and error handling |