Configure DNS routing policies and health checks

This page describes how to configure DNS routing policies, and how to enable health checking by using Cloud DNS. Before you use this page, familiarize yourself with the DNS policies overview.

Permissions required for this task

To perform this task, you must have been granted the following permissions or the following IAM roles.

Permissions

dns.networks.useHealthSignals to create a policy with health checks for internal passthrough Network Load Balancers
dns.changes.create to create a resource record set
dns.managedZones.list to list managed zones

Roles

roles/dns.admin
roles/dns.peer

To use DNS routing policies, create a resource record set and choose one of the following DNS routing policies to apply to the resource record set:

DNS routing policies also support multiple IP addresses for each geographic location. When specified for a given geographic location, multiple IP addresses are returned according to an equal weight WRR policy. Combining̦ a geo-based routing policy with a custom-weighted WRR policy is not supported.

Only one type of routing policy can be applied to a resource record set at a time. You cannot combine routing policies except when configuring a failover routing policy, in which case, you can set a geolocation routing policy as the backup. Global access must be enabled for regional load balancers.

Create DNS routing policies for private zones

Before you create DNS routing policies for private zones, complete the following steps.

Create a private zone.
Set up one of the following internal load balancers:
Create forwarding rules for the internal load balancer.
Set up health checking for the internal load balancer.

To create a resource record set and apply a routing policy to it, follow these steps.

Console

Start the configuration

In the Google Cloud console, go to the Cloud DNS zones page.

Go to Cloud DNS zones
Click the name of the managed zone that you want to add the record to.
On the Zone details page, click Add with routing policy.

Base data

On the Create record set with routing policy page, in the DNS name field, enter the subdomain of the DNS zone—for example, mail. The trailing dot is automatically added at the end.

Note: Adding the @ symbol in this field causes the record to fail.
Select the Resource record type—for example, A.
In the TTL field, enter a numeric value for the resource record's time to live, which is the amount of time that it can be cached. This value must be a positive integer.
Optional: In the TTL unit menu, select the unit of time—for example, minutes.
Click Next.

Routing policy type

In the Routing policy list, select Weighted round robin, Geolocation, or Failover.
Click Next.

Routing policy data

If you selected Weighted round robin, in the Weighted round robin policy routing data section, do the following:
1. In the Weight field, enter the weight corresponding to this subsection of the resource record (RR) data. This weight should be a nonnegative number from 0.0 to 1000.0. Ratio of traffic routed to the target is calculated from the ratio of individual weight over the total across all weights.
2. In the Add health checked target section, do the following:
  1. In the Project list, select the project where the forwarding rule exists.
  2. In the Forwarding rule list, select a forwarding rule.
    
    The forwarding rule specifies an internal IP address, port, and a regional backend service or an HTTP(S) proxy. For Cloud DNS to work with health checks, you must enable global access for the internal load balancer.
3. To allow IPv4 addresses without health checking, select Allow IPv4 addresses without health checking.
4. In the IPv4 Address field, enter an IPv4 address.
If you selected Geolocation, do the following:
1. For Geo fencing, select Disabled or Enabled. Enabling geofencing restricts the traffic to a specific geolocation even if all the endpoints in that geolocation are unhealthy.
2. In the Source region menu, select a valid Google Cloud source region, such as asia-east1.
3. In the Add health checked target section, do the following:
  1. In the Project list, select the project where the forwarding rule exists.
  2. In the Type list, select internal passthrough Network Load Balancer, internal Application Load Balancer, or cross-region internal Application Load Balancer.
  3. In the Forwarding rule list, select a forwarding rule.
    
    The forwarding rule specifies an internal IP address, port, and a regional backend service or an HTTP(S) proxy. For Cloud DNS to work with health checks, you must enable global access for the internal load balancer.
4. To allow IPv4 addresses without health checking, select Allow IPv4 addresses without health checking.
5. In the IPv4 Address field, enter an IPv4 address.
If you selected Failover, do the following:

Note: You can configure a failover routing policy only for private zones.
1. In the Trickle traffic (%) field, enter the percentage of the traffic sent to the failover targets, regardless of the health check status of the primary targets.
2. In the Primary targets section, do the following:
  1. In the Project list, select the project where the forwarding rule exists.
  2. In the Type list, select internal passthrough Network Load Balancer, internal Application Load Balancer, or cross-region internal Application Load Balancer.
  3. In the Forwarding rule list, select a forwarding rule.
    
    The forwarding rule specifies an internal IP address, port, and a regional backend service or an HTTP(S) proxy. For Cloud DNS to work with health checks, you must enable global access for the internal load balancer.
3. In the Backup geolocation policy section, do the following:
  1. For Geo fencing, select Disabled or Enabled. Enabling geofencing restricts the traffic to a specific geolocation even if all the endpoints in that geolocation are unhealthy.
  2. In the Source region menu, select a valid Google Cloud source region, such as asia-east1.
  3. In the Add health checked target section, do the following:
    1. In the Project list, select the project where the forwarding rule exists.
    2. In the Type list, select internal passthrough Network Load Balancer, internal Application Load Balancer, or cross-region internal Application Load Balancer.
    3. In the Forwarding rule list, select a forwarding rule.
    When all primary IP addresses are unhealthy, traffic is automatically handled according to the backup geolocation policy.
4. To allow IPv4 addresses without health checking, select Allow IPv4 addresses without health checking.
5. In the IPv4 Address field, enter an IPv4 address.
Click Next.

Review and create

Click Review.
Review your Cloud DNS record set with routing policy configuration.
Optional: Click Equivalent comment line to view the gcloud CLI command to create this record set with routing policy.
Click Create.

gcloud

A ResourceRecordSet can contain a routingPolicy or rrdatas, but not both. Changing between rrdatas or routingPolicy is supported when updating ResourceRecordSets. For example, if you want to update a ResourceRecordSet that contains rrdatas, you can delete the rrdatas and add a routingPolicy to the same ResourceRecordSet.

To create a ResourceRecordSet and apply a routing policy to it, follow these steps.

Run the gcloud dns record-sets create command:

For GEO policies

gcloud dns record-sets create RRSET_NAME \
    --ttl=TTL \
    --type=RRSET_TYPE \
    --zone=MANAGED_ZONE \
    --routing-policy-type=GEO \
    --routing-policy-data=ROUTING_POLICY_DATA \
    --enable-health-checking

For WRR policies

gcloud dns record-sets create RRSET_NAME \
    --ttl=TTL \
    --type=RRSET_TYPE \
    --zone=MANAGED_ZONE \
    --routing-policy-type=WRR \
    --routing-policy-data=ROUTING_POLICY_DATA \
    --enable-health-checking

For geofenced policies

gcloud dns record-sets create RRSET_NAME \
    --ttl=TTL \
    --type=RRSET_TYPE \
    --zone=MANAGED_ZONE \
    --routing-policy-type=GEO \
    --routing-policy-data=ROUTING_POLICY_DATA \
    --enable-geo-fencing
    --enable-health-checking

For failover policies

gcloud dns record-sets create RRSET_NAME \
    --ttl=TTL \
    --type=RRSET_TYPE \
    --zone=MANAGED_ZONE \
    --routing-policy-type=FAILOVER \
    --enable-geo-fencing \
    --routing-policy-primary-data=ROUTING_POLICY_PRIMARY_DATA \
    --routing-policy-backup-data-type=ROUTING_POLICY_BACKUP_DATA_TYPE \
    --routing-policy-backup-data=ROUTING_POLICY_BACKUP_DATA \
    --backup-data-trickle-ratio=BACKUP_DATA_TRICKLE_RATIO \
    --enable-health-checking

Replace the following:

RRSET_NAME: the DNS name that matches the incoming queries with this zone's DNS name as its suffix, such as service.example.com
TTL: the TTL, in seconds, in which the resolver caches ResourceRecordSet, such as 30
RRSET_TYPE: the resource record type of this ResourceRecordSet, such as A

For a list of supported record types, see Select resource record types.
MANAGED_ZONE: the managed zone that this ResourceRecordSet is affiliated with, such as service-zone. The name of this ResourceRecordSet must have the DNS name of the managed zone as its suffix
ROUTING_POLICY_TYPE: the type of routing policy.

Enter WRR for weighted round robin, GEO for geo-location, or FAILOVER for failover policies. You cannot modify this field after a policy has a chosen type; you can only delete the policy and add a new policy with the different type.
ROUTING_POLICY_DATA: the routing policy data
- For --routing-policy-type=WRR, enter a semicolon-delimited list in the format ${weight_percent}:${rrdatas}, such as .8=203.0.113.1;.2=198.51.100.1. Specify the weight as a nonnegative decimal. The ratio of traffic routed to the target is calculated from the ratio of individual weight over the total across all weights. Forwarding rule names are acceptable values and result in health checking.
- For --routing-policy-type=GEO, enter a semicolon-delimited list in the format ${region}=${IP_address}, such as asia-east1=198.51.100.1;us-central1=203.0.113.1. You can specify multiple IP addresses for a single region by adding IP addresses separated by a comma. Forwarding rule names are acceptable values and result in health checking.
- For --routing-policy-type=FAILOVER, enter the name of the forwarding rule that you created in the format ${region}=${Forwarding rule name}.
  
  Note: When health checks are disabled, the rrdata for WRR or GEO can be RFC 1035 format DNS rrdata.
  When health checks are enabled, the rrdata is a reference to a forwarding rule. Cloud DNS uses the IP address of the forwarding rule as the rrdata. Cloud DNS monitors the health status of the forwarding rule and removes the forwarding rule from the results if it becomes unhealthy.
You must specify both the routing policy type and the routing policy data. If you specify one, you cannot leave the other flag unspecified.
--enable-geo-fencing: for GEO routing policies, this determines whether traffic should failover across regions if all endpoints in a region are unhealthy. When set, Cloud DNS always directs queries to the nearest region, even if all endpoints in that region are unhealthy. Use --no-enable-geo-fencing to disable geofencing. When not set, Cloud DNS directs queries to the next nearest region when all endpoints in a region are unhealthy. This defaults to false.
ROUTING_POLICY_PRIMARY_DATA: the primary target to use for FAILOVER routing policies. This target must be a reference to one or more forwarding rules, such as forwarding-rule-1. As long as at least one of these forwarding rules is healthy, the IP addresses of all healthy forwarding rules are used to answer queries for this name.
ROUTING_POLICY_BACKUP_DATA: the backup target to use for FAILOVER routing policies. These targets are used when all forwarding rules specified in --routing-policy-primary-data are unhealthy. Cloud DNS only supports geo-based backup targets. The format of this field matches that of --routing-policy-data when --routing-policy-type = 'GEO', such as asia-east1=forwarding-rule-2.
ROUTING_POLICY_BACKUP_DATA_TYPE: for FAILOVER routing policies, the type of routing policy the backup data uses. This must be GEO.
BACKUP_DATA_TRICKLE_RATIO: the ratio of traffic to send to the backup targets, even when the primaries are healthy. The ratio must be between 0 and 1, such as 0.1. The default is set to 0.
--enable-health-checking: the flag to enable health checking. When you use this flag, you must provide the forwarding rule name instead of the IP address in the --routing-policy-data field.

API

Use the resourceRecordSets.create method.

For geolocation policies

POST https://www.googleapis.com/dns/v1/projects/PROJECT_ID/managedZones/MANAGED_ZONE/rrsets
{
            "name": "RRSET_NAME",
            "type": "RRSET_TYPE",
            "ttl": TTL,
            "routingPolicy": {
          "geo": {
              "items": [
              {
                  "location": "LOCATION",
                  "healthCheckedTargets": {
                     "internalLoadBalancers": [
                      {
                       "loadBalancerType": "LOAD_BALANCER_TYPE"
                       "ipAddress": "IP_ADDRESS"
                       "port" : "PORT_NUMBER"
                       "ipProtocol": "IP_PROTOCOL"
                       "networkUrl": "NETWORK_URL"
                       "project": "PROJECT"
                       "region": "REGION"
                      }
                     ]
                  }
              },
              {
                  "location": "LOCATION",
                  "healthCheckedTargets": {
                     "internalLoadBalancers": [
                      {
                       "loadBalancerType": "LOAD_BALANCING_TYPE"
                       "ipAddress": "IP_ADDRESS"
                       "port" : "PORT_NUMBER"
                       "ipProtocol": "IP_PROTOCOL"
                       "networkUrl": "NETWORK_URL"
                       "project": "PROJECT"
                       "region": "REGION"
                      }
                     ]
                  }
              },
              }
           ]

        }
     }
}

For WRR policies

POST https://www.googleapis.com/dns/v1/projects/PROJECT_ID/managedZones/MANAGED_ZONE/rrsets
{
  "name": "RRSET_NAME",
        "type": "RRSET_TYPE",
        "ttl": TTL,
  "routingPolicy": {
    "wrr": {
      "items": [
        {
          "weight": WEIGHT,
          "healthCheckedTargets": {
            "internalLoadBalancers": [
              {
                "loadBalancerType": "LOAD_BALANCER_TYPE"
                "ipAddress": "IP_ADDRESS"
                "port" : "PORT_NUMBER"
                "ipProtocol": "IP_PROTOCOL"
                "networkUrl": "NETWORK_URL"
                "project": "PROJECT"
                "region": "REGION"
              }
            ]
          }
        },
        {
          "weight": WEIGHT,
          "healthCheckedTargets": {
            "internalLoadBalancers": [
              {
                "loadBalancerType": "LOAD_BALANCER_TYPE"
                "ipAddress": "IP_ADDRESS"
                "port" : "PORT_NUMBER"
                "ipProtocol": "IP_PROTOCOL"
                "networkUrl": "NETWORK_URL"
                "project": "PROJECT"
                "region": "REGION"
              }
            ]
          }
        },
      ]
    }
  }
}

For failover for geolocation policies

In the failover option, Cloud DNS only supports GEO policies.

POST https://www.googleapis.com/dns/v1/projects/PROJECT_ID/managedZones/MANAGED_ZONE/rrsets
{
  "name": "RRSET_NAME",
        "type": "RRSET_TYPE",
        "ttl": TTL,
  "routingPolicy": {
    "primaryBackup": {
      "trickleTraffic": TRICKLE_TRAFFIC,
      "primaryTargets": {
        "internalLoadBalancers": [
          {
            "ipAddress": "IP_ADDRESS"
            "ipProtocol": "IP_PROTOCOL"
            "loadBalancerType": "LOAD_BALANCER_TYPE"
            "networkUrl": "NETWORK_URL"
            "port": "PORT_NUMBER"
            "project": "PROJECT"
            "region": "REGION"
           }
         ]
       },
       "backupGeoTargets": {
         "enableFencing": ENABLE_FENCING,
         "items": [
           {
             "location": "LOCATION",
             "rrdatas": [
               "RRDATA"
             ]
           },
           {
             "location": "LOCATION",
             "rrdatas": [
               "RRDATA"
             ]
           }
         ]
       }
     },
   }
}

Replace the following:

PROJECT_ID: the ID of the project
MANAGED_ZONE: the managed zone that this ResourceRecordSet is affiliated with, such as service-zone; the name of this ResourceRecordSet must have the DNS name of the managed zone as its suffix
RRSET_NAME: the DNS name that matches the incoming queries with this zone's DNS name as its suffix, such as service.example.com
RRSET_TYPE: the resource record type of this ResourceRecordSet, such as A
TTL: the TTL, in seconds, in which the resolver caches this ResourceRecordSet, such as 30
TRICKLE_TRAFFIC: the ratio of traffic to send to the backup targets even when the primaries are healthy; the ratio must be between 0 and 1, such as 0.1
ENABLE_FENCING: for GEO routing policies, this determines whether traffic should failover across regions if all endpoints in a region are unhealthy. When set, Cloud DNS always directs queries to the nearest region, even if all endpoints in that region are unhealthy. When not set, Cloud DNS directs queries to the next nearest region when all endpoints in a region are unhealthy. This defaults to false.
LOCATION: for GEO policies, the geolocation for which you need to create the policy, such as asia-east1
WEIGHT: for WRR policies, a semicolon-delimited list in the format ${weight_percent}=${rrdatas}, such as .8=10.128.1.1;.2=10.130.1.1; specify the weight as any nonnegative decimal
Note: You must specify weight as a nonnegative number. The ratio of traffic routed to the target is calculated from the ratio of individual weight over the total across all weights.
RRDATA: an arbitrary value associated with the resource record set, such as 198.51.100.5; you can also enter multiple values, rrdata1 rrdata2 rrdata3, such as 198.51.100.1 203.0.113.1...
LOAD_BALANCER_TYPE: the type of load balancer, such as regionalL4ilb, globalL7ilb, or regionalL7ilb. This setting is optional.
IP_ADDRESS: the IP address that the forwarding rule serves
PORT_NUMBER: the port number
IP_PROTOCOL: defines the protocol used for the health check; valid options are tcp and udp
NETWORK_URL: the network URL to which this forwarding rule applies
REGION: the region in which you created the forwarding rule

Create DNS routing policies for public zones (Preview)

To create a resource record set and apply a routing policy to it, follow these steps.

gcloud

To enable health checking in DNS routing policies for public zones, you must first create a health check for external endpoints.

Run the gcloud beta compute health-checks create command:
```
gcloud beta compute health-checks create PROTOCOL HEALTH_CHECK_NAME \
    --global
    --check-interval=CHECK_INTERVAL \
    --source-regions=SOURCE_REGIONS \
    --port=PORT_NUMBER
```
Replace the following:
- PROTOCOL: the protocol used for the health check. Valid options are http, https, ssl, or tcp.
- HEALTH_CHECK_NAME: the name of the health check.
- CHECK_INTERVAL: the amount of time from the start of one health check probe system's connection to the start of the next one. Units are seconds. The CHECK_INTERVAL value must be between 30 and 300 seconds. If omitted, Google Cloud uses a value of 30 (30 seconds).
- SOURCE_REGIONS: a comma-separated list of Google Cloud regions from which you want to send health check probes.
- PORT_NUMBER: the port number for health check requests.
To create a ResourceRecordSet and apply a routing policy to it, follow these steps.

Run the gcloud beta dns record-sets create command:
WRR
```
gcloud beta dns record-sets create RRSET_NAME \
    --ttl=TTL \
    --type=RRSET_TYPE \
    --zone=MANAGED_ZONE \
    --routing-policy-type=WRR \
    --routing-policy-data=ROUTING_POLICY_DATA \
    --enable-health-checking \
    --health-check=HEALTH_CHECK_NAME
```
Replace the following:
- RRSET_NAME: the DNS name that matches the incoming queries with this zone's DNS name as its suffix, such as service.example.com.
- TTL: the TTL, in seconds, in which the resolver caches this ResourceRecordSet, such as 30.
- RRSET_TYPE: the resource record type of this ResourceRecordSet, such as A.
- MANAGED_ZONE: the managed zone that this ResourceRecordSet is affiliated with, such as service-zone. The name of this ResourceRecordSet must have the DNS name of the managed zone as its suffix.
- ROUTING_POLICY_DATA: the routing policy data. Enter a semicolon-delimited list in the format ${weight_percent}:${rrdatas}, such as .8=203.0.113.1;.2=198.51.100.1. Specify the weight as a nonnegative decimal. Weight must be a nonnegative number from 0to 1000.
- HEALTH_CHECK_NAME: the name of the health check that you created in the previous step.
  
  You must specify both the routing policy type and the routing policy data. If you specify one, you cannot leave the other flag unspecified.
Geolocation
```
gcloud beta dns record-sets create RRSET_NAME \
    --ttl=TTL \
    --type=RRSET_TYPE \
    --zone=MANAGED_ZONE \
    --routing-policy-type=GEO \
    --routing-policy-data=ROUTING_POLICY_DATA \
    --enable-health-checking \
    --health-check=HEALTH_CHECK_NAME
```
Replace the following:
- RRSET_NAME: the DNS name that matches the incoming queries with this zone's DNS name as its suffix, such as service.example.com.
- TTL: the TTL, in seconds, in which the resolver caches this ResourceRecordSet, such as 30.
- RRSET_TYPE: the resource record type of this ResourceRecordSet, such as A.
- MANAGED_ZONE: the managed zone that this ResourceRecordSet is affiliated with, such as service-zone. The name of this ResourceRecordSet must have the DNS name of the managed zone as its suffix.
- ROUTING_POLICY_DATA: the routing policy data. Enter a semicolon-delimited list in the format ${region}=${IP_address},${IP_address}, such as asia-east1=198.51.100.1;us-central1=203.0.113.1, 203.0.113.2. You can specify multiple IP addresses for a single region by adding IP addresses separated by a comma. Forwarding rule names are acceptable values and result in health checking.
- HEALTH_CHECK_NAME: the name of the health check that you created in the previous step.
  
  You must specify both the routing policy type and the routing policy data. If you specify one, you cannot leave the other flag unspecified.
Geolocation with geofence
```
gcloud beta dns record-sets create RRSET_NAME \
    --ttl=TTL \
    --type=RRSET_TYPE \
    --zone=MANAGED_ZONE \
    --routing-policy-type=GEO \
    --routing-policy-data=ROUTING_POLICY_DATA \
    --enable-geo-fencing
    --enable-health-checking
    --health-check=HEALTH_CHECK_NAME
```
Replace the following:
- RRSET_NAME: the DNS name that matches the incoming queries with this zone's DNS name as its suffix, such as service.example.com.
- TTL: the TTL, in seconds, in which the resolver caches this ResourceRecordSet, such as 30.
- RRSET_TYPE: the resource record type of this ResourceRecordSet, such as A.
- MANAGED_ZONE: the managed zone that this ResourceRecordSet is affiliated with, such as service-zone. The name of this ResourceRecordSet must have the DNS name of the managed zone as its suffix.
- ROUTING_POLICY_DATA: the routing policy data. Enter a semicolon-delimited list in the format ${region}=${IP_address}, such as asia-east1=198.51.100.1;us-central1=203.0.113.1. You can specify multiple IP addresses for a single region by adding IP addresses separated by a comma. Forwarding rule names are acceptable values and result in health checking.
- HEALTH_CHECK_NAME: the name of the health check that you created in the previous step.
  
  --enable-geo-fencing: for GEO routing policies, this determines whether traffic should failover across regions if all endpoints in a region are unhealthy. When set, Cloud DNS always directs queries to the nearest region, even if all endpoints in that region are unhealthy. Use --no-enable-geo-fencing to disable geofencing. When not set, Cloud DNS directs queries to the next nearest region when all endpoints in a region are unhealthy. This defaults to false.
  
  --enable-health-checking: the flag to enable health checking. When you use this flag, you must provide the forwarding rule name instead of the IP address in the --routing-policy-data field.
  
  You must specify both the routing policy type and the routing policy data. If you specify one, you cannot leave the other flag unspecified.
Failover
```
gcloud beta dns record-sets create RRSET_NAME \
    --ttl=TTL \
    --type=RRSET_TYPE \
    --zone=MANAGED_ZONE \
    --routing-policy-type=FAILOVER \
    --routing-policy-primary-data=ROUTING_POLICY_PRIMARY_DATA \
    --routing-policy-backup-data=ROUTING_POLICY_BACKUP_DATA \
    --routing-policy-backup-data-type=ROUTING_POLICY_BACKUP_DATA_TYPE \
    --backup-data-trickle-ratio=BACKUP_DATA_TRICKLE_RATIO \
    --enable-health-checking
    --health-check=HEALTH_CHECK_NAME
```
Replace the following:
- RRSET_NAME: the DNS name that matches the incoming queries with this zone's DNS name as its suffix, such as service.example.com.
- TTL: the TTL, in seconds, in which the resolver caches this ResourceRecordSet, such as 30
- RRSET_TYPE: the resource record type of this ResourceRecordSet, such as A.
- MANAGED_ZONE: the managed zone that this ResourceRecordSet is affiliated with, such as service-zone. The name of this ResourceRecordSet must have the DNS name of the managed zone as its suffix.
- ROUTING_POLICY_PRIMARY_DATA: the primary target to use for FAILOVER routing policies. This target must be a reference to one or more forwarding rules, such as forwarding-rule-1. As long as at least one of these forwarding rules is healthy, the IP addresses of all healthy forwarding rules are used to answer queries for this name.
- ROUTING_POLICY_BACKUP_DATA: the backup target to use for FAILOVER routing policies. These targets are used when all forwarding rules specified in --routing-policy-primary-data are unhealthy. Cloud DNS only supports geo-based backup targets. The format of this field matches that of --routing-policy-data when --routing-policy-type = 'GEO', such as asia-east1=forwarding-rule-2.
- ROUTING_POLICY_BACKUP_DATA_TYPE: for FAILOVER routing policies, the type of routing policy the backup data uses. This must be GEO.
- BACKUP_DATA_TRICKLE_RATIO: the ratio of traffic to send to the backup targets, even when the primaries are healthy. The ratio must be between 0 and 1, such as 0.1. The default is set to 0.
- HEALTH_CHECK_NAME: the name of the health check that you created in the previous step.
  
  You must specify both the routing policy type and the routing policy data. If you specify one, you cannot leave the other flag unpopulated.

API

A ResourceRecordSet resource can contain either a routingPolicy or an rrdatas section, but not both. You can switch between rrdata and routingPolicy when you update a ResourceRecordSets resource. For example, if you want to update a ResourceRecordSet resource that contains an rrdata section, you can delete the rrdatas section and add a routingPolicy section to the same ResourceRecordSet resource.

To enable health checking in DNS routing policies for public zones, you must first create an internet-based health check.

Use the healthChecks.insert method.
To create a ResourceRecordSet and apply a routing policy to it, follow these steps.

Use the resourceRecordSets.create method.
WRR
```
    POST https://www.googleapis.com/dns/v1/projects/PROJECT_ID/managedZones/MANAGED_ZONE/rrsets
    {
        "name": "RRSET_NAME",
        "type": "RRSET_TYPE",
        "ttl": TTL,
        "routingPolicy": {
            "healthCheck": "https://www.googleapis.com/compute/beta/projects/PROJECT_ID/global/healthChecks/HEALTH_CHECK_NAME"
            "wrr": {
                "items": [{
                    "weight": WEIGHT,
                    "healthCheckedTargets": {
                        "rrdata": ["RRDATA"]
                    }
                }, {
                    "weight": 1.0,
                    "healthCheckedTargets": {
                        "rrdata": ["RRDATA", "RRDATA"]
                    }
                }]
            }
        }
    }
  
```
Replace the following:
- PROJECT_ID: the ID of the project
- MANAGED_ZONE: the managed zone that this ResourceRecordSet is affiliated with, such as service-zone; the name of this ResourceRecordSet must have the DNS name of the managed zone as its suffix.
- RRSET_NAME: the DNS name that matches the incoming queries with this zone's DNS name as its suffix, such as service.example.com.
- TTL: the TTL, in seconds, in which the resolver caches this ResourceRecordSet, such as 30.
- RRSET_TYPE: the resource record type of this ResourceRecordSet, such as A.
- HEALTH_CHECK_NAME: the name of the health check.
- WEIGHT: for WRR policies, a semicolon-delimited list in the format ${weight_percent}=${rrdatas}, such as .8=10.128.1.1;.2=10.130.1.1; specify the weight as any nonnegative decimal. Note: You must specify weight as a nonnegative number. The ratio of traffic routed to the target is calculated from the ratio of individual weight over the total across all weights.
- RRDATA: an arbitrary value associated with the resource record set, such as 198.51.100.5; you can also enter multiple values, rrdata1,rrdata2,rrdata3, such as 198.51.100.1, 203.0.113.1.
Geolocation
```
    POST https://www.googleapis.com/dns/v1/projects/PROJECT_ID/managedZones/MANAGED_ZONE/rrsets
    {
        "name": "RRSET_NAME",
        "type": "RRSET_TYPE",
        "ttl": TTL,
        "routingPolicy": {
            "healthCheck": "https://www.googleapis.com/compute/beta/projects/PROJECT_ID/global/healthChecks/HEALTH_CHECK_NAME"
            "geo": {
          "enableFencing": ENABLE_FENCING
                "items": [{
                    "location": "LOCATION",
                    "healthCheckedTargets": {
                        "rrdata": ["RRDATA"]
                    }
                }, {
                    "location": "LOCATION",
                    "healthCheckedTargets": {
                        "rrdata": ["RRDATA", "RRDATA"]
                    }
                }]
            }
        }
    }
  
```
Replace the following:
- PROJECT_ID: the ID of the project
- MANAGED_ZONE: the managed zone that this ResourceRecordSet is affiliated with, such as service-zone; the name of this ResourceRecordSet must have the DNS name of the managed zone as its suffix.
- RRSET_NAME: the DNS name that matches the incoming queries with this zone's DNS name as its suffix, such as service.example.com.
- RRSET_TYPE: the resource record type of this ResourceRecordSet, such as A.
- TTL: the TTL, in seconds, in which the resolver caches this ResourceRecordSet, such as 30.
- HEALTH_CHECK_NAME: the name of the health check.
- ENABLE_FENCING: for GEO routing policies, this determines whether traffic should failover across regions if all endpoints in a region are unhealthy. When set, Cloud DNS always directs queries to the nearest region, even if all endpoints in that region are unhealthy. When not set, Cloud DNS directs queries to the next nearest region when all endpoints in a region are unhealthy. Valid options are true and false. The default setting for this is false.
- LOCATION: for GEO policies, the geolocation for which you need to create the policy, such as asia-east1.
- RRDATA: an arbitrary value associated with the resource record set, such as 198.51.100.5; you can also enter multiple values, rrdata1,rrdata2,rrdata3, such as 198.51.100.1, 203.0.113.1.
Failover
In the failover option, Cloud DNS only supports GEO policies.
```
    POST https://www.googleapis.com/dns/v1/projects/PROJECT_ID/managedZones/MANAGED_ZONE/rrsets
    {
        "name": "RRSET_NAME",
        "type": "RRSET_TYPE",
        "ttl": TTL,
        "routingPolicy": {
            "healthCheck": "https://www.googleapis.com/compute/beta/projects/PROJECT_ID/global/healthChecks/HEALTH_CHECK_NAME"
            "primaryBackup": {
                "trickleTraffic": TRICKLE_TRAFFIC,
                "primaryTargets": {
                    "rrdata": ["RRDATA"]
                }
                "backupGeoTargets": {
                    "enableFencing": ENABLE_FENCING,
                    "items": [{
                        "location": "LOCATION",
                        "rrdatas": ["RRDATA]
                    }, {
                        "location": "LOCATION",
                        "rrdatas": ["RRDATA", "RRDATA"]
                    }]
                }
            }
        }
    }
  
```
Replace the following:
- PROJECT_ID: the ID of the project
- MANAGED_ZONE: the managed zone that this ResourceRecordSet is affiliated with, such as service-zone; the name of this ResourceRecordSet must have the DNS name of the managed zone as its suffix.
- RRSET_NAME: the DNS name that matches the incoming queries with this zone's DNS name as its suffix, such as service.example.com.
- RRSET_TYPE: the resource record type of this ResourceRecordSet, such as A.
- TTL: the TTL, in seconds, in which the resolver caches this ResourceRecordSet, such as 30.
- HEALTH_CHECK_NAME: the name of the health check.
- TRICKLE_TRAFFIC: the ratio of traffic to send to the backup targets even when the primaries are healthy; the ratio must be between 0 and 1, such as 0.1
- RRDATA: an arbitrary value associated with the resource record set, such as 198.51.100.5; you can also enter multiple values, rrdata1,rrdata2,rrdata3, such as 198.51.100.1, 203.0.113.1.
- ENABLE_FENCING: for GEO routing policies, this determines whether traffic must failover across regions if all endpoints in a region are unhealthy. When set, Cloud DNS always directs queries to the nearest region, even if all endpoints in that region are unhealthy. When not set, Cloud DNS directs queries to the next nearest region when all endpoints in a region are unhealthy. The default setting for this is false.
- LOCATION: for GEO policies, the geolocation for which you need to create the policy, such as asia-east1.

Update DNS routing policies

To update a resource record set's routing policy, follow these steps.

Console

In the Google Cloud console, go to the Cloud DNS zones page.

Go to Cloud DNS zones
Click the zone for which you want to update the resource record set's routing policy.
On the Zone details page, next to the resource record set that you want to update, click Edit.
After you have made the updates, click Save.

gcloud

Run the gcloud dns record-sets update command:

WRR

gcloud dns record-sets update RRSET_NAME \
  --ttl=TTL \
  --type=RRSET_TYPE \
  --zone=MANAGED_ZONE \
  --routing-policy-type=WRR \
  --routing-policy-data=ROUTING_POLICY_DATA \
  --enable-health-checking

Geolocation

gcloud dns record-sets update RRSET_NAME \
  --ttl=TTL \
  --type=RRSET_TYPE \
  --zone=MANAGED_ZONE \
  --routing-policy-type=GEO \
  --routing-policy-data=ROUTING_POLICY_DATA \
  --enable-health-checking

Geolocation with geofence

gcloud dns record-sets update RRSET_NAME \
  --ttl=TTL \
  --type=RRSET_TYPE \
  --zone=MANAGED_ZONE \
  --routing-policy-type=GEO \
  --routing-policy-data=ROUTING_POLICY_DATA \
  --enable-geo-fencing
  --enable-health-checking

Failover

gcloud dns record-sets update RRSET_NAME \
  --ttl=TTL \
  --type=RRSET_TYPE \
  --zone=MANAGED_ZONE \
  --routing-policy-type=FAILOVER \
  --enable-geo-fencing \
  --routing-policy-primary-data=ROUTING_POLICY_PRIMARY_DATA \
  --routing-policy-backup-data=ROUTING_POLICY_BACKUP_DATA \
  --backup-data-trickle-ratio=BACKUP_DATA_TRICKLE_RATIO \
  --enable-health-checking

Replace the following:

RRSET_NAME: the DNS name that matches the incoming queries with this zone's DNS name as its suffix, such as service.example.com
TTL: the TTL, in seconds, in which the resolver caches this ResourceRecordSet, such as 30
RRSET_TYPE: the resource record type of this ResourceRecordSet, such as A

For a list of supported record types, see Select resource record types.
MANAGED_ZONE: the managed zone that this ResourceRecordSet is affiliated with, such as service-zone. The name of this ResourceRecordSet must have the DNS name of the managed zone as its suffix
ROUTING_POLICY_TYPE: the type of routing policy.

Enter WRR for weighted round robin, GEO for geo-location, or FAILOVER for failover policies. You cannot modify this field after a policy has a chosen type; you can only delete the policy and add a new policy with the different type.
ROUTING_POLICY_DATA: the routing policy data
- For --routing-policy-type=WRR, enter a semicolon-delimited list in the format ${weight_percent}:${rrdatas}, such as .8=203.0.113.1;.2=198.51.100.1. Specify the weight as a nonnegative decimal. The ratio of traffic routed to the target is calculated from the ratio of individual weight over the total across all weights. Forwarding rule names are acceptable values and result in health checking.
- For --routing-policy-type=GEO, enter a semicolon-delimited list in the format ${region}=${IP_address}, such as asia-east1=198.51.100.1;us-central1=203.0.113.1. You can specify multiple IP addresses for a single region by adding IP addresses separated by a comma. Forwarding rule names are acceptable values and result in health checking.
- For --routing-policy-type=FAILOVER, enter the name of the forwarding rule that you created in the format ${region}=${Forwarding rule name}.
  
  Note: When health checks are disabled, the rrdata for `WRR` or `GEO` can be RFC 1035 format DNS `rrdata`.
  When health checks are enabled, the rrdata is a reference to a forwarding rule. Cloud DNS uses the IP address of the forwarding rule as the rrdata. If it becomes unhealthy, Cloud DNS monitors the health status of the forwarding rule and removes the forwarding rule from the results.
You must specify both the routing policy type and the routing policy data. If you specify one, you cannot leave the other flag unpopulated.
--enable-geo-fencing: for GEO routing policies, this determines whether traffic should fail over across regions if all endpoints in a region are unhealthy. When set, Cloud DNS always directs queries to the nearest region, even if all endpoints in that region are unhealthy. Use --no-enable-geo-fencing to disable geofencing. When unset, all endpoints in a region are unhealthy and Cloud DNS directs queries to the next nearest region. The default setting for this is false.
ROUTING_POLICY_PRIMARY_DATA: the primary target to use for FAILOVER routing policies. This target must be a reference to one or more forwarding rules, such as forwarding-rule-1. As long as at least one of these forwarding rules is healthy, the IP addresses of all healthy forwarding rules are used to answer queries for this name.
ROUTING_POLICY_BACKUP_DATA: the backup target to use for FAILOVER routing policies. These targets are used when all forwarding rules specified in --routing-policy-primary-data are unhealthy. Cloud DNS only supports geo-based backup targets. The format of this field matches that of --routing-policy-data when --routing-policy-type = 'GEO', such as asia-east1=forwarding-rule-2.
BACKUP_DATA_TRICKLE_RATIO: the ratio of traffic to send to the backup targets even when the primaries are healthy. The ratio must be between 0 and 1, such as 0.1. The default is set to 0.
--enable-health-checking: Enables the health checking of forwarding rules that are provided as rrdata to --routing-policy-data.

API

Use the resourceRecordSets.patch method. Specify only one of rrset.rrdatas or rrset.routingPolicy. If specifying routingPolicy, you must specify the new routingPolicy field in its entirety.

WRR

For WRR policies, use the following method:

PATCH https://www.googleapis.com/dns/v1/projects/PROJECT_ID/managedZones/MANAGED_ZONE/rrsets
{
      "name": "RRSET_NAME.",
      "type": "RRSET_TYPE",
      "ttl": TTL,
      "routingPolicy": {
        "wrrPolicy": {
            "item": [
                  {
                      "weight": WEIGHT,
                      "rrdatas": ["RRDATA"]
                  },
                  {
                      "weight": WEIGHT,
                      "rrdatas": ["RRDATA"]
                  }
            ],
          }
    }
}

Geolocation

Use the following method:

PATCH https://www.googleapis.com/dns/v1/projects/PROJECT_ID/managedZones/MANAGED_ZONE/rrsets
{
          "name": "RRSET_NAME",
          "type": "RRSET_TYPE",
          "ttl": TTL,
          "routingPolicy": {
        "geo": {
            "items": [
            {
                "location": "LOCATION",
                "healthCheckedTargets": {
                  "internalLoadBalancers": [
                    {
                    "loadBalancerType": "LOAD_BALANCER_TYPE"
                    "ipAddress": "IP_ADDRESS"
                    "port" : "PORT_NUMBER"
                    "ipProtocol": "IP_PROTOCOL"
                    "networkUrl": "NETWORK_URL"
                    "project": "PROJECT"
                    "region": "REGION"
                    }
                  ]
                }
            },
            {
                "location": "LOCATION",
                "healthCheckedTargets": {
                  "internalLoadBalancers": [
                    {
                    "loadBalancerType": "LOAD_BALANCING_TYPE"
                    "ipAddress": "IP_ADDRESS"
                    "port" : "PORT_NUMBER"
                    "ipProtocol": "IP_PROTOCOL"
                    "networkUrl": "NETWORK_URL"
                    "project": "PROJECT"
                    "region": "REGION"
                    }
                  ]
                }
            },
            }
        ]

      }
  }
}

Replace the following:

PROJECT_ID: the ID of the project
MANAGED_ZONE: the managed zone that this ResourceRecordSet is affiliated with, such as service-zone; the name of this ResourceRecordSet must have the DNS name of the managed zone as its suffix
RRSET_NAME: the DNS name that matches the incoming queries with this zone's DNS name as its suffix, such as service.example.com
RRSET_TYPE: the resource record type of this ResourceRecordSet, such as A
TTL: the TTL, in seconds, in which the resolver caches this ResourceRecordSet, such as 30
TRICKLE_TRAFFIC: the ratio of traffic to send to the backup targets even when the primaries are healthy; the ratio must be between 0 and 1, such as 0.1
ENABLE_FENCING: for GEO routing policies, this determines whether traffic should failover across regions if all endpoints in a region are unhealthy. When set, Cloud DNS always directs queries to the nearest region, even if all endpoints in that region are unhealthy. When not set, Cloud DNS directs queries to the next nearest region when all endpoints in a region are unhealthy. This defaults to false.
LOCATION: for GEO policies, the geolocation for which you need to update the policy, such as asia-east1
WEIGHT: for WRR policies, a semicolon-delimited list in the format ${weight_percent}=${rrdatas}, such as .8=10.128.1.1;.2=10.130.1.1; specify the weight as any nonnegative decimal
Note: You must specify weight as a nonnegative number. The ratio of traffic routed to the target is calculated from the ratio of individual weight over the total across all weights.
RRDATA: an arbitrary value associated with the resource record set, such as 198.51.100.5; you can also enter multiple values, rrdata1 rrdata2 rrdata3, such as 198.51.100.1 203.0.113.1...
LOAD_BALANCER_TYPE: the type of load balancer, such as regionalL4ilb, globalL7ilb, or regionalL7ilb. This setting is optional.
IP_ADDRESS: the IP address that the forwarding rule serves
PORT_NUMBER: the port number
IP_PROTOCOL: defines the protocol used for the health check; valid options are tcp and udp
NETWORK_URL: the network URL to which this forwarding rule applies
REGION: the region in which you created the forwarding rule

Delete DNS routing policies

To delete a routing policy, you must delete the resource record set that contains the routing policy. To do so, follow these steps.

Console

In the Google Cloud console, go to the Cloud DNS zones page.

Go to Cloud DNS zones
Click the zone for which you want to delete the resource record set.
On the Zone details page, next to the DNS name of the resource record set that you want to delete, select the checkbox.
Click Delete record sets.

gcloud

Run the gcloud dns record-sets delete command:

gcloud dns record-sets delete RRSET_NAME \
    --type=RRSET_TYPE \
    --zone=MANAGED_ZONE \

Replace the following:

RRSET_NAME: the DNS name that matches the incoming queries with this zone's DNS name as its suffix, such as service.example.com
RRSET_TYPE: the resource record type of this ResourceRecordSet, such as A

For a list of supported record types, see Selecting resource record types.
MANAGED_ZONE: the managed zone that this ResourceRecordSet is affiliated with, such as service-zone; the name of this ResourceRecordSet must have the DNS name of the managed zone as its suffix

API

Use the resourceRecordSets.delete method:

DELETE https://www.googleapis.com/dns/v1/projects/PROJECT_ID/managedZones/MANAGED_ZONE/rrsets/RRSET_NAME/RRSET_TYPE

Replace the following:

PROJECT_ID: the ID of the project
MANAGED_ZONE: the managed zone that this ResourceRecordSet is affiliated with, such as my-zone-name; the name of this ResourceRecordSet must have the DNS name of the managed zone as its suffix
RRSET_NAME: the DNS name that matches the incoming queries with this zone's DNS name as its suffix, such as test.example.com
RRSET_TYPE: the resource record type of this ResourceRecordSet, such as A

What's next

To work with managed zones, see Create, modify, and delete zones.
To find solutions for common issues that you might encounter when using Cloud DNS, see Troubleshooting.
To get an overview of Cloud DNS, see Cloud DNS overview.