11 / 67 · 06 Observability-Driven Testing · Canary Deployments← prev⊞ allnext →☰ Read as one page
2.5Key Decisions for Canary Deployments
How Much Traffic to Send to Canary?
| Traffic Percentage | Use Case | Risk Level |
|---|---|---|
| 1% | High-risk changes (payment, auth) | Very low |
| 5% | Standard feature releases | Low |
| 10% | Low-risk changes with high confidence | Low |
| 25% | Changes that need more traffic volume for statistical significance | Medium |
How Long to Observe?
The observation window depends on traffic volume and the statistical significance required:
- High traffic services (>1000 rps): 10-15 minutes provides enough data points
- Medium traffic (100-1000 rps): 30-60 minutes
- Low traffic (<100 rps): 2-6 hours (consider synthetic traffic augmentation)
What Metrics to Compare?
At minimum, compare these between canary and baseline:
- Error rate (critical -- always include)
- Latency percentiles (p50, p95, p99)
- Saturation metrics (CPU, memory per pod)
- Business metrics (conversion rate, revenue per request -- if available in real-time)