Облачная платформаAdvanced

Configuring Network Policies to Restrict Pod Access

Эта статья полезна?

Язык статьи: Английский

Network policies are designed by Kubernetes to restrict pod access. Like a firewall at the application layer, network policies enhance network security. The capabilities supported by network policies depend on the capabilities of the network add-ons of the cluster.

By default, if a namespace does not have any policy, pods in the namespace accept traffic from any source and send traffic to any destination.

The following selectors are available for network policies:

namespaceSelector: selects particular namespaces for which all pods should be allowed as ingress sources or egress destinations.

podSelector: selects particular pods in the same namespace as the network policy which should be allowed as ingress sources or egress destinations.

ipBlock: selects particular CIDR blocks to allow as ingress sources or egress destinations.

Relationships Between Network Policies and Cluster Types

Cluster Type	CCE Standard Cluster	CCE Standard Cluster	CCE Turbo Cluster
Network Model	Tunnel Network	VPC Network	Cloud Native Network 2.0
Network policies	Enabled by default	Disabled by default (To use network policies, enable DataPlane V2 when creating a cluster.)	Disabled by default (To use network policies, enable DataPlane V2 when creating a cluster.)
Data plane implementation	OpenvSwitch	eBPF	eBPF
Cluster versions for ingress rules	All versions	v1.27.16-r30, v1.28.15-r20, v1.29.13-r0, v1.30.10-r0, v1.31.6-r0, or later	Clusters v1.34.3-r10 or later
Cluster versions for egress rules	v1.23 and later		Clusters v1.34.3-r10 or later
Selector for ingress rules	namespaceSelector podSelector	namespaceSelector podSelector ipBlock	namespaceSelector podSelector ipBlock
Selector for egress rules	namespaceSelector podSelector ipBlock	namespaceSelector podSelector ipBlock	namespaceSelector podSelector ipBlock
Supported OS	EulerOS CentOS HCE OS 2.0	HCE OS 2.0 is supported. Clusters v1.28.15-r70, v1.29.15-r30, v1.30.14-r30, v1.31.10-r30, v1.32.6-r30, v1.33.5-r20, v1.34.1-r0, and later versions support Ubuntu 22.04.	HCE OS 2.0 is supported. Clusters v1.28.15-r70, v1.29.15-r30, v1.30.14-r30, v1.31.10-r30, v1.32.6-r30, v1.33.5-r20, v1.34.1-r0, and later versions support Ubuntu 22.04.
IPv6 network policies	Not supported	Not supported	Supported
Secure containers	Not supported	Not supported	Not supported
IPBlock scope	Not limited	Subnets within the pod CIDR block, Service CIDR block, and node IP addresses	Subnets within the pod CIDR block, Service CIDR block, and node IP addresses
Limiting ClusterIP access through workload labels	Not supported	Supported	Supported
Limiting the internal cloud server CIDR block of 100.125.0.0/16	Supported	Supported	Supported
SCTP	Not supported	Supported	Supported
Always allowing access to pods on a node from other nodes	Supported	Supported	Supported
Configuring EndPort in network policies	Not supported	Supported	Supported

Note

If you upgrade a CCE standard cluster with a tunnel network to a version that supports egress rules in in-place mode, the rules will not work because the node OS is not upgraded. In this case, reset the node.

When a network policy is enabled for a cluster with a tunnel network, a pod's source IP address is embedded in the optional field of packets it sends to any Service CIDR block. This enables the configuration of network policy rules on the destination pod, taking into account the source IP address of the pod.

Using Ingress Rules Through YAML

Scenario 1: Configure a network policy to allow only pods with specified labels to access the target pod.

Figure 1 podSelector

The pod labeled with role=db only permits access to its port 6379 from pods labeled with role=frontend. To achieve this, take the following steps:

Create the access-ingress1.yaml file.

vim access-ingress1.yaml

File content:

apiVersion: networking.k8s.io/v1
kind: NetworkPolicy
metadata:
  name: access-ingress1
  namespace: default
spec:
  podSelector:                  # The rule applies only to pods labeled with  role=db .
    matchLabels:
      role: db
  ingress:                      # This is an ingress rule.
  - from:
    - podSelector:              # Allow access only from pods labeled with  role=frontend .
        matchLabels:
          role: frontend
    ports:                      # Only TCP can be used to access port 6379.
    - protocol: TCP
      port: 6379

Run the following command to create the network policy defined in the access-ingress1.yaml file:

kubectl apply -f access-ingress1.yaml

Expected output:

networkpolicy.networking.k8s.io/access-ingress1 created

Scenario 2: Configure a network policy to allow only pods in a specific namespace to access the target pod.

Figure 2 namespaceSelector

The pod labeled with role=db only permits access to its port 6379 from pods in the namespace labeled with project=myproject. To achieve this, take the following steps:

Create the access-ingress2.yaml file.

vim access-ingress2.yaml

File content:

apiVersion: networking.k8s.io/v1
kind: NetworkPolicy
metadata:
  name: access-ingress2
spec:
  podSelector:                  # The rule applies only to pods labeled with  role=db .
    matchLabels:
      role: db
  ingress:                      # This is an ingress rule.
  - from:
    - namespaceSelector:        # Allow access only from pods in namespaces labeled with  project=myproject .
        matchLabels:
          project: myproject
    ports:                      # Only TCP can be used to access port 6379.
    - protocol: TCP
      port: 6379

Run the following command to create the network policy defined in the access-ingress2.yaml file:

kubectl apply -f access-ingress2.yaml

Expected output:

networkpolicy.networking.k8s.io/access-ingress2 created

Scenario 3: Configure a network policy to allow only pods with specific labels in a specific namespace to access the target pod.

Figure 3 Using both podSelector and namespaceSelector

The pod labeled with role=db only allows access to its port 6379 from pods with label role=frontend in the namespace labeled with project=myproject. To achieve this, take the following steps:

Create the access-ingress3.yaml file.

vim access-ingress3.yaml

File content:

apiVersion: networking.k8s.io/v1
kind: NetworkPolicy
metadata:
  name: access-ingress3
spec:
  podSelector:                  # The rule applies only to pods labeled with  role=db .
    matchLabels:
      role: db
  ingress:                      # This is an ingress rule.
  - from:
    - namespaceSelector:        # Allow access only from pods in namespaces labeled with  project=myproject .
        matchLabels:
          project: myproject
      podSelector:              # Allow access only from pods labeled with  role=frontend .
        matchLabels:
          role: frontend
    ports:                      # Only TCP can be used to access port 6379.
    - protocol: TCP
      port: 6379

Run the following command to create the network policy defined in the access-ingress3.yaml file:

kubectl apply -f access-ingress3.yaml

Expected output:

networkpolicy.networking.k8s.io/access-ingress3 created

Using Egress Rules Through YAML

Scenario 1: Controlled by a preset network policy, pods can only access specific addresses.

Figure 4 ipBlock

The pods labeled role=db only allow access to 172.16.0.0/16, excluding 172.16.0.40/32. To achieve this, take the following steps:

Create the access-egress1.yaml file.

vim access-egress1.yaml

File content:

apiVersion: networking.k8s.io/v1
kind: NetworkPolicy
metadata:
  name: access-egress1
  namespace: default
spec:
  policyTypes:                  # This policy type must be specified for egress rules.
    - Egress
  podSelector:                  # The rule applies only to pods labeled with  role=db .
    matchLabels:
      role: db
  egress:                       # This is an egress rule.
  - to:
    - ipBlock:
        cidr: 172.16.0.0/16    # Allows access to this CIDR block in the outbound direction.
        except:
        - 172.16.0.40/32        # Blocks access to this CIDR block, which is in the range specified by the  cidr  parameter.

Run the following command to create the network policy defined in the access-egress1.yaml file:

kubectl apply -f access-egress1.yaml

Expected output:

networkpolicy.networking.k8s.io/access-egress1 created

Scenario 2: Controlled by a preset network policy, a pod can only be accessed by pods with specific labels, while this pod itself can only access specific pods.

Figure 5 Using both ingress and egress

The pod labeled with role=db only permits access to its port 6379 from pods labeled with role=frontend, and this pod can only access the pods labeled with role=web. You can use the same rule to configure both ingress and egress in a network policy. To achieve this, take the following steps:

Create the access-egress2.yaml file.

vim access-egress2.yaml

File content:

apiVersion: networking.k8s.io/v1
kind: NetworkPolicy
metadata:
  name: access-egress2
  namespace: default
spec:
  policyTypes:
  - Ingress
  - Egress
  podSelector:                  # The rule applies only to pods labeled with  role=db .
    matchLabels:
      role: db
  ingress:                      # This is an ingress rule.
  - from:
    - podSelector:              # Allow access only from pods labeled with  role=frontend .
        matchLabels:
          role: frontend
    ports:                      # Only TCP can be used to access port 6379.
    - protocol: TCP
      port: 6379
  egress:                       # This is an egress rule.
  - to:
    - podSelector:              # The rule takes effect for pods with the  role=web  label.
        matchLabels:
          role: web

Run the following command to create the network policy defined in the access-egress2.yaml file:

kubectl apply -f access-egress2.yaml

Expected output:

networkpolicy.networking.k8s.io/access-egress2 created

Creating a Network Policy on the Console

Choose Policies in the navigation pane, click the Network Policies tab, and click Create Network Policy in the upper right corner.

Policy Name: Specify a network policy name.

Namespace: Select a namespace in which the network policy is applied.

Selector: Enter a label, select the pod to be associated, and click Add. You can also click Reference Workload Label to use the label of an existing workload.

Inbound Rule: Click

to add an inbound rule. For details about parameter settings, see Table 1.

Table 1 Adding an inbound rule
Parameter	Description
Protocol & Port	Select the protocol type and port. Currently, TCP and UDP are supported.
Source CIDR Block	For clusters v1.27.16-r10, v1.28.15-r0, v1.29.10-r0, v1.30.6-r0, or later versions with DataPlane V2 enabled, you can configure the source CIDR block. The specified source CIDR block allows traffic from a destination CIDR block (multiple exception CIDR blocks can be specified). Separate the destination and exception CIDR blocks using a vertical bar (\|). If there are multiple exception CIDR blocks, separate them using commas (,). For example, 172.17.0.0/16\|172.17.1.0/24,172.17.2.0/24 indicates that 172.17.0.0/16 is accessible, but 172.17.1.0/24 and 172.17.2.0/24 are inaccessible.
Source Namespace	Select a namespace whose objects can be accessed. If this parameter is not specified, the object belongs to the same namespace as the current policy.
Source Pod Label	Allow accessing the pods with this label. If this parameter is not specified, all pods in the namespace can be accessed.

Outbound Rule: Click

to add an outbound rule. For details about parameter settings, see Table 2.

Table 2 Adding an outbound rule
Parameter	Description
Protocol & Port	Select the protocol type and port. Currently, TCP and UDP are supported. If this parameter is not specified, the protocol type is not limited.
Destination CIDR Block	Allow requests to be routed to a specified CIDR block (and not to the exception CIDR blocks). Separate the destination and exception CIDR blocks using a vertical bar (\|). If there are multiple exception CIDR blocks, separate them using commas (,). For example, 172.17.0.0/16\|172.17.1.0/24,172.17.2.0/24 indicates that 172.17.0.0/16 is accessible, but 172.17.1.0/24 and 172.17.2.0/24 are inaccessible.
Destination Namespace	Select a namespace whose objects can be accessed. If this parameter is not specified, the object belongs to the same namespace as the current policy.
Destination Pod Label	Allow accessing the pods with this label. If this parameter is not specified, all pods in the namespace can be accessed.

Click OK.

Parent topic: Pod Network Settings

Предыдущая статья

Configuring QoS for a Pod

Следующая статья

DataPlane V2

Эта статья полезна?

Поддержка Юридические документы Политика конфиденциальности