Incorrect hostname for VMs with multiple network interfaces

[d3c3balus]

Given 2 VPCs:

NAME                    REGION        NETWORK                     RANGE
secondary-subnet-first  europe-west1  vpc-secondary-europe-west1  172.16.0.0/16
subnet-first            europe-west1  vpc-primary-europe-west1    10.1.0.0/16

I create multiple VMs, each with a network connection in each VPC.

NAME         ZONE            MACHINE_TYPE  PREEMPTIBLE  INTERNAL_IP          EXTERNAL_IP  STATUS
k8s-master   europe-west1-b  e2-medium                  10.1.0.2,172.16.0.2               RUNNING
k8s-worker1  europe-west1-b  e2-medium                  10.1.0.3,172.16.0.3               RUNNING
k8s-worker2  europe-west1-b  e2-medium                  10.1.0.4,172.16.0.4               RUNNING

Although, nic0 points to vpc-primary-europe-west1 (subnet 10.1.0.0/16), the /etc/hosts file contains wrong IP (from the secondary VPC, nic1):

# cat /etc/hosts
127.0.0.1 localhost

# The following lines are desirable for IPv6 capable hosts
::1 ip6-localhost ip6-loopback
fe00::0 ip6-localnet
ff00::0 ip6-mcastprefix
ff02::1 ip6-allnodes
ff02::2 ip6-allrouters
ff02::3 ip6-allhosts
169.254.169.254 metadata.google.internal metadata
172.16.0.2 k8s-master.europe-west1-b.c.k8snet.internal k8s-master  # Added by Google
169.254.169.254 metadata.google.internal  # Added by Google

Note the row marked as "Added by Google".

Additional info: the default route is correctly using nic0 (primary VPC):

default via 10.1.0.1 dev ens4 proto dhcp src 10.1.0.2 metric 100 
10.1.0.0/16 via 10.1.0.1 dev ens4 proto dhcp src 10.1.0.2 metric 100 
10.1.0.1 dev ens4 proto dhcp scope link src 10.1.0.2 metric 100 
172.16.0.0/16 via 172.16.0.1 dev ens5 
172.16.0.1 dev ens5 scope link 

Conclusion: GCP incorrectly sets the hostname for the secondary IP instead of the IP of nic0.
Impact: some tools (e.g. Kubelet) uses the IP of the hostname of the node and incorrectly uses the secondary IP.