BackendServiceFailoverPolicy(
mapping=None, *, ignore_unknown_fields=False, **kwargs
)Applicable only to Failover for Internal TCP/UDP Load Balancing and Network Load Balancing. On failover or failback, this field indicates whether connection draining will be honored. GCP has a fixed connection draining timeout of 10 minutes. A setting of true terminates existing TCP connections to the active pool during failover and failback, immediately draining traffic. A setting of false allows existing TCP connections to persist, even on VMs no longer in the active pool, for up to the duration of the connection draining timeout (10 minutes).
Attributes | |
|---|---|
| Name | Description |
disable_connection_drain_on_failover |
bool
This can be set to true only if the protocol is TCP. The default is false. |
drop_traffic_if_unhealthy |
bool
Applicable only to Failover for Internal TCP/UDP Load Balancing and Network Load Balancing, If set to true, connections to the load balancer are dropped when all primary and all backup backend VMs are unhealthy.If set to false, connections are distributed among all primary VMs when all primary and all backup backend VMs are unhealthy. The default is false. |
failover_ratio |
float
Applicable only to Failover for Internal TCP/UDP Load Balancing and Network Load Balancing. The value of the field must be in the range [0, 1]. If the value is 0, the load balancer performs a failover when the number of healthy primary VMs equals zero. For all other values, the load balancer performs a failover when the total number of healthy primary VMs is less than this ratio. |