Whitelist IP addresses

cli/cmd/lib_realtime_apis.go

@@ -167,6 +167,9 @@ func describeModelInput(status *status.Status, predictor *userconfig.Predictor,
 	if predictor.Type == userconfig.TensorFlowPredictorType && !cachingEnabled {
 		apiTFLiveReloadingSummary, err := getAPITFLiveReloadingSummary(apiEndpoint)
 		if err != nil {
+			if strings.Contains(errors.Message(err), "context deadline exceeded") {
+				return "error retrieving the models' metadata schema: unable to connect to the API, you either do not have access or the API is too busy" + "\n"
+			}
 			return "error retrieving the models' metadata schema: " + errors.Message(err) + "\n"
 		}
 		t, err := parseAPITFLiveReloadingSummary(apiTFLiveReloadingSummary)
@@ -178,6 +181,9 @@ func describeModelInput(status *status.Status, predictor *userconfig.Predictor,
 
 	apiModelSummary, err := getAPIModelSummary(apiEndpoint)
 	if err != nil {
+		if strings.Contains(errors.Message(err), "context deadline exceeded") {
+			return "error retrieving the models' metadata schema: unable to connect to the API, you either do not have access or the API is too busy" + "\n"
+		}
 		return "error retrieving the models' metadata schema: " + errors.Message(err) + "\n"
 	}
 	t, err := parseAPIModelSummary(apiModelSummary)
@@ -196,7 +202,7 @@ func getModelFromModelID(modelID string) (modelName string, modelVersion int64,
 
 func makeRequest(request *http.Request) (http.Header, []byte, error) {
 	client := http.Client{
-		Timeout: 600 * time.Second,
+		Timeout: 5 * time.Second,
 		Transport: &http.Transport{
 			TLSClientConfig: &tls.Config{InsecureSkipVerify: true},
 		},

docs/clusters/management/create.md

@@ -82,6 +82,12 @@ operator_load_balancer_scheme: internet-facing
 #   - availability_zone: us-west-2b
 #     subnet_id: subnet-0faed05adf6042ab7
 
+# restrict access to APIs by cidr blocks/ip address ranges
+api_load_balancer_cidr_white_list: [0.0.0.0/0]
+
+# restrict access to the Operator by cidr blocks/ip address ranges
+operator_load_balancer_cidr_white_list: [0.0.0.0/0]
+
 # additional tags to assign to AWS resources (all resources will automatically be tagged with cortex.dev/cluster-name: <cluster_name>)
 tags:  # <string>: <string> map of key/value pairs
 

docs/clusters/networking/load-balancers.md

@@ -2,8 +2,8 @@
 
 ![api architecture diagram](https://user-images.githubusercontent.com/808475/103417256-dd6e9700-4b3e-11eb-901e-90425f1f8fd4.png)
 
-All APIs share a single API load balancer. By default, the API load balancer is public. You can configure your API load balancer to be private by setting `api_load_balancer_scheme: internal` in your cluster configuration file (before creating your cluster). This will make your API only accessible through [VPC Peering](vpc-peering.md).
+All APIs share a single API load balancer. By default, the API load balancer is public. You can configure your API load balancer to be private by setting `api_load_balancer_scheme: internal` in your cluster configuration file (before creating your cluster). This will make your API only accessible through [VPC Peering](vpc-peering.md). You can enforce that incoming requests to APIs must originate from specific ip address ranges by specifying `api_load_balancer_cidr_white_list: [<CIDR list>]` in your cluster configuration.
 
 The SSL certificate on the API load balancer is autogenerated during installation using `localhost` as the Common Name (CN). Therefore, clients will need to skip certificate verification when making HTTPS requests to your APIs (e.g. `curl -k https://***`), or make HTTP requests instead (e.g. `curl http://***`). Alternatively, you can enable HTTPS by using a [custom domain](custom-domain.md) or by [creating an API Gateway](https.md) to forward requests to your API load balancer.
 
-There is a separate load balancer for the Cortex operator. By default, the operator load balancer is public. You can configure your operator load balancer to be private by setting `operator_load_balancer_scheme: internal` in your cluster configuration file (before creating your cluster). You can use [VPC Peering](vpc-peering.md) to enable your Cortex CLI to connect to your cluster operator from another VPC.
+There is a separate load balancer for the Cortex operator. By default, the operator load balancer is public. You can configure your operator load balancer to be private by setting `operator_load_balancer_scheme: internal` in your cluster configuration file (before creating your cluster). You can use [VPC Peering](vpc-peering.md) to enable your Cortex CLI to connect to your cluster operator from another VPC. You can enforce that incoming requests to the Cortex operator must originate from specific ip address ranges by specifying `operator_load_balancer_cidr_white_list: [<CIDR list>]` in your cluster configuration.

manager/manifests/istio.yaml.j2

@@ -50,6 +50,9 @@ spec:
           service:
             type: LoadBalancer
             externalTrafficPolicy: Cluster # https://medium.com/pablo-perez/k8s-externaltrafficpolicy-local-or-cluster-40b259a19404, https://www.asykim.com/blog/deep-dive-into-kubernetes-external-traffic-policies
+            {% if config.get('operator_load_balancer_cidr_white_list', [])|length > 0 %}
+            loadBalancerSourceRanges: {{ config['operator_load_balancer_cidr_white_list'] }}
+            {% endif %}
             selector:
               app: operator-istio-gateway
               istio: ingressgateway-operator
@@ -107,6 +110,9 @@ spec:
             {% endif %}
           service:
             type: LoadBalancer
+            {% if config.get('api_load_balancer_cidr_white_list', [])|length > 0 %}
+            loadBalancerSourceRanges: {{ config['api_load_balancer_cidr_white_list'] }}
+            {% endif %}
             externalTrafficPolicy: Cluster # https://medium.com/pablo-perez/k8s-externaltrafficpolicy-local-or-cluster-40b259a19404, https://www.asykim.com/blog/deep-dive-into-kubernetes-external-traffic-policies
             selector:
               app: apis-istio-gateway

pkg/types/clusterconfig/cluster_config.go

@@ -98,18 +98,20 @@ type CoreConfig struct {
 }
 
 type ManagedConfig struct {
-	NodeGroups                 []*NodeGroup       `json:"node_groups" yaml:"node_groups"`
-	Tags                       map[string]string  `json:"tags" yaml:"tags"`
-	AvailabilityZones          []string           `json:"availability_zones" yaml:"availability_zones"`
-	SSLCertificateARN          *string            `json:"ssl_certificate_arn,omitempty" yaml:"ssl_certificate_arn,omitempty"`
-	IAMPolicyARNs              []string           `json:"iam_policy_arns" yaml:"iam_policy_arns"`
-	SubnetVisibility           SubnetVisibility   `json:"subnet_visibility" yaml:"subnet_visibility"`
-	Subnets                    []*Subnet          `json:"subnets,omitempty" yaml:"subnets,omitempty"`
-	NATGateway                 NATGateway         `json:"nat_gateway" yaml:"nat_gateway"`
-	APILoadBalancerScheme      LoadBalancerScheme `json:"api_load_balancer_scheme" yaml:"api_load_balancer_scheme"`
-	OperatorLoadBalancerScheme LoadBalancerScheme `json:"operator_load_balancer_scheme" yaml:"operator_load_balancer_scheme"`
-	VPCCIDR                    *string            `json:"vpc_cidr,omitempty" yaml:"vpc_cidr,omitempty"`
-	CortexPolicyARN            string             `json:"cortex_policy_arn" yaml:"cortex_policy_arn"` // this field is not user facing
+	NodeGroups                        []*NodeGroup       `json:"node_groups" yaml:"node_groups"`
+	Tags                              map[string]string  `json:"tags" yaml:"tags"`
+	AvailabilityZones                 []string           `json:"availability_zones" yaml:"availability_zones"`
+	SSLCertificateARN                 *string            `json:"ssl_certificate_arn,omitempty" yaml:"ssl_certificate_arn,omitempty"`
+	IAMPolicyARNs                     []string           `json:"iam_policy_arns" yaml:"iam_policy_arns"`
+	SubnetVisibility                  SubnetVisibility   `json:"subnet_visibility" yaml:"subnet_visibility"`
+	Subnets                           []*Subnet          `json:"subnets,omitempty" yaml:"subnets,omitempty"`
+	NATGateway                        NATGateway         `json:"nat_gateway" yaml:"nat_gateway"`
+	APILoadBalancerScheme             LoadBalancerScheme `json:"api_load_balancer_scheme" yaml:"api_load_balancer_scheme"`
+	OperatorLoadBalancerScheme        LoadBalancerScheme `json:"operator_load_balancer_scheme" yaml:"operator_load_balancer_scheme"`
+	APILoadBalancerCIDRWhiteList      []string           `json:"api_load_balancer_cidr_white_list,omitempty" yaml:"api_load_balancer_cidr_white_list,omitempty"`
+	OperatorLoadBalancerCIDRWhiteList []string           `json:"operator_load_balancer_cidr_white_list,omitempty" yaml:"operator_load_balancer_cidr_white_list,omitempty"`
+	VPCCIDR                           *string            `json:"vpc_cidr,omitempty" yaml:"vpc_cidr,omitempty"`
+	CortexPolicyARN                   string             `json:"cortex_policy_arn" yaml:"cortex_policy_arn"` // this field is not user facing
 }
 
 type NodeGroup struct {
@@ -663,6 +665,34 @@ var ManagedConfigStructFieldValidations = []*cr.StructFieldValidation{
 			return LoadBalancerSchemeFromString(str), nil
 		},
 	},
+	{
+		StructField: "APILoadBalancerCIDRWhiteList",
+		StringListValidation: &cr.StringListValidation{
+			Validator: func(addresses []string) ([]string, error) {
+				for i, address := range addresses {
+					_, err := validateCIDR(address)
+					if err != nil {
+						return nil, errors.Wrap(err, fmt.Sprintf("index %d", i))
+					}
+				}
+				return addresses, nil
+			},
+		},
+	},
+	{
+		StructField: "OperatorLoadBalancerCIDRWhiteList",
+		StringListValidation: &cr.StringListValidation{
+			Validator: func(addresses []string) ([]string, error) {
+				for i, address := range addresses {
+					_, err := validateCIDR(address)
+					if err != nil {
+						return nil, errors.Wrap(err, fmt.Sprintf("index %d", i))
+					}
+				}
+				return addresses, nil
+			},
+		},
+	},
 	{
 		StructField: "OperatorLoadBalancerScheme",
 		StringValidation: &cr.StringValidation{
@@ -676,7 +706,7 @@ var ManagedConfigStructFieldValidations = []*cr.StructFieldValidation{
 	{
 		StructField: "VPCCIDR",
 		StringPtrValidation: &cr.StringPtrValidation{
-			Validator: validateVPCCIDR,
+			Validator: validateCIDR,
 		},
 	},
 }
@@ -1083,7 +1113,7 @@ func validateBucketName(bucket string) (string, error) {
 	return bucket, nil
 }
 
-func validateVPCCIDR(cidr string) (string, error) {
+func validateCIDR(cidr string) (string, error) {
 	_, _, err := net.ParseCIDR(cidr)
 	if err != nil {
 		return "", errors.WithStack(err)