Observability
OpenTelemetry + Langfuse + Jaeger tracing
Avg Latency
1.7s
-12% vs last week
Error Rate
0.4%
Stable
Throughput
142/min
+18% vs last week
Active Traces
34
right now
Model Performance
| Model | Calls (30d) | Avg Latency | P99 | Error Rate | Tokens | Cost |
|---|---|---|---|---|---|---|
| GPT-4o | 45.2k | 1.8s | 4.5s | 0.3% | 12.4M | $892 |
| Claude Sonnet 4.6 | 23.1k | 2.1s | 5.2s | 0.5% | 8.2M | $567 |
| Gemini 3 Pro | 8.4k | 1.5s | 3.8s | 0.8% | 3.1M | $234 |
| GPT-4o-mini | 32.7k | 0.8s | 2.1s | 0.1% | 18.9M | $156 |