ARC2 Proxy
Load balancing Routing rule integer · seconds

health_check_interval

How often each backend is probed.

Config key
health_check_interval
Scope
Routing rule
Type
integer · seconds
Default
none

Seconds between health probes. Lower means faster detection of a failed backend but more probe traffic; higher is quieter but slower to react.

Picking a cadence
5–15s is a sensible range for most pools. Going below a couple of seconds rarely helps and adds noise to your origins.
health_check_interval = 10