Total requests: 0
👤
Your frontend
sends requests to a single URL
/p/abc123/
⚖️
WatsonLB
routes · queues · heals
Backend A
Render
● healthy
0 req · 100% up
Backend B
Railway
● healthy
0 req · 100% up
Backend C
Fly.io
◌ cold-starting
0 req · 72% up
Backend D
Koyeb
● healthy
0 req · 98% up

Click a server to toggle it down/up

healthy
cold
down
recovering
packet in flight
activity log
← Simulation started. Click "Send request" or watch auto-fire.
Round-robin
Each request goes to the next healthy backend. If one fails, it is skipped automatically and the circuit breaker trips.
Request queue
When all backends are cold-starting or down, requests are held in-memory and flushed the moment any backend responds.
🔁
Auto-recovery
WatsonLB pings all backends continuously. When a backend recovers, it is automatically re-added to the rotation.
Set this up for your backends →