Health Checks and Automatic Failover

Why Health Checks Matter

Without health checks, traffic goes to dead servers. Health checks detect failures and remove unhealthy origins automatically.

How It Works

Periodic HTTP requests to your health path. 200 = healthy. 5xx/timeout = unhealthy.

Setting

Default

|---------|---------|

Interval	30 seconds
Path	/health
Timeout	5 seconds
Unhealthy threshold	2 failures
Healthy threshold	2 successes

Failover

Unhealthy origin removed → traffic redistributed. Origin recovers → added back automatically.

Best Practices

1. Keep health checks lightweight (don't query DB)

2. Use threshold of 2+ to avoid false positives

3. Test failover regularly

4. Monitor health check logs