Advanced Features
Load Balancing
Distribute API traffic across multiple models and providers to maximize throughput, minimize latency, and avoid single-provider bottlenecks — all configured in the Dashboard.
Load Balancing Strategies
OneRouter supports multiple routing strategies, configurable per channel in the Dashboard:
| Strategy | How It Works | Best For |
|---|---|---|
| Weighted Random | Distributes requests based on assigned weights (e.g., 70% Model A, 30% Model B) | Gradual migrations, A/B testing, cost optimization |
| Priority (Failover) | Routes to the highest-priority available model; falls back down the chain | Production resilience, primary/backup setups |
| Rate-Limit Based | Distributes overflow traffic when a model hits its rate limit | High-throughput batch processing |
Configuring Load Balancing in the Dashboard
- Log in to the OneRouter Dashboard at api.onerouter.app/dashboard
- Navigate to Channels in the sidebar
- For weighted distribution: create multiple channels for the same model group, assign different weights to each
- For priority-based: configure model priority ordering within a single channel
- Save the configuration — changes take effect immediately with no restart required
Example: Weighted Distribution Across Providers
To split traffic 60/40 between two providers for the same model type:
| Channel | Model | Weight | Share |
|---|---|---|---|
| Channel A | GPT-4o (OpenAI) | 60 | 60% of traffic |
| Channel B | Claude Opus 4.8 (Anthropic) | 40 | 40% of traffic |
Users call one model name — OneRouter handles the distribution transparently. No client-side changes needed.
Pro tip: Start with a 90/10 split when testing a new provider. Gradually increase the weight as you gain confidence in reliability and response quality.