Health Checks
Active probes remove failing backends from rotation and restore them only after recovery thresholds are met.
Portfolio Signal
Infrastructure Depth In One Running System
Round Robin
Fair request distribution across healthy nodes for homogeneous backend pools.
Least Connections
Routes new traffic toward less-busy backends when request durations are uneven.
Circuit Breakers
Stops repeatedly failing upstreams from absorbing traffic during an outage window.
Rate Limiting
Token-bucket protection reduces overload risk before traffic reaches upstream services.
Metrics Dashboard
Live counters expose request volume, retries, failovers, upstream errors, and backend selection.
Live Control Plane
Routing Strategy And Backend Health
Reads from /admin/backends, /admin/strategy, and /admin/cost.
Current Strategy
Loading...
Proxy Endpoint
/proxy/
Estimated Cost
loading...
Metrics Dashboard
Reliability And Traffic Counters
Status Classes
Backend Selection
| Route | Method | Samples | Avg Latency |
|---|---|---|---|
| Loading metrics... | |||
Request Path
How EdgeBalancer Handles Traffic
- Client request reaches
/proxy/*through CloudFront, ALB, and ECS/Fargate. - Rate limiting and request tracing run before proxy forwarding.
- EdgeBalancer chooses a healthy backend with the active strategy.
- Retries and circuit breakers isolate upstream failures from client traffic.
- Metrics and structured logs record request, backend, latency, and failover behavior.
Routing Algorithms
Strategy Trade-offs
Round Robin
Best when backend capacity and response time are similar.
Least Connections
Best when some requests hold connections longer than others.
Weighted Routing
Best when stronger nodes should carry a larger traffic share.
Consistent Hashing
Best when cache locality or request affinity matters.
AI Copilot
Runtime-Aware Architecture Guidance
AI Provider
loading...
AI response will appear here.