Introducing Container Network Logs with Advanced Container Networking Services for AKS
Overview of container network logs Container network logs offer a comprehensive way to monitor network traffic in AKS clusters. Two modes of support, stored-logs and on-demand logs, provides debugging flexibility with cost optimization. The on-demand mode provides a snapshot of logs with queries and visualization with Hubble CLI UI for specific scenarios and does not use log storage to persist the logs. The stored-logs mode when enabled continuously collects and persists logs based on user-defined filters. Logs can be stored either in Azure Log Analytics (managed) or locally (unmanaged). Managed storage: Logs are forwarded to Azure Log Analytics for secure, scalable, and compliant storage. This enables advanced analytics, anomaly detection, and historical trend analysis. Both basic and analytics table plans are supported for storage. Unmanaged storage: Logs are stored locally on the host nodes under /var/log/acns/hubble. These logs are rotated automatically at 50 MB to manage storage efficiently. These logs can be exported to external logging systems or collectors for further analysis. Use cases Connectivity monitoring: Identify and visualize how Kubernetes workloads communicate within the cluster and with external endpoints, helping to resolve application connectivity issues efficiently. Troubleshooting network errors: Gain deep granular visibility into dropped packets, misconfigurations, or errors with details on where and why errors are occurring (TCP/UDP, DNS, HTTP) for faster root cause analysis. Security policy enforcement: Detect and analyze suspicious traffic patterns to strengthen cluster security and ensure regulatory compliance. How it works Container network logs use eBPF technology with Cilium to capture network flows from AKS nodes. Log collection is disabled by default. Users can enable log collection by defining custom resources (CRs) to specify the types of traffic to monitor, such as namespaces, pods, services, or protocols. The Cilium agent collects and processes this traffic, storing logs in JSON format. These logs can either be retained locally or integrated with Azure Monitoring for long-term storage and advanced analytics and visualization with Azure managed Grafana. Fig1: Container network logs overview If using managed storage, users will enable Azure monitor log collection using Azure CLI or ARM templates. Here’s a quick example of enabling container network logs on Azure monitor using the CLI: az aks enable-addons -a monitoring --enable-high-log-scale-mode -g $RESOURCE_GROUP -n $CLUSTER_NAME az aks update --enable-acns \ --enable-retina-flow-logs \ -g $RESOURCE_GROUP \ -n $CLUSTER_NAME Key benefits Faster issue resolution: Detailed logs enable quick identification of connectivity and performance issues. Operational efficiency: Advanced filtering reduces data management overhead. Enhanced application reliability: Proactive monitoring ensures smoother operations. Cost optimization: Customized logging scopes minimize storage and data ingestion costs. Streamlined compliance: Comprehensive logs support audits and security requirements. Observing logs in Azure managed Grafana dashboards Users can visualize container network logs in Azure managed Grafana dashboards, which simplify monitoring and analysis: Flow logs dashboard: View internal communication between Kubernetes workloads. This dashboard highlights metrics such as total requests, dropped packets, and error rates. Error logs dashboard: Easily zoom in only on the logs which show errors for faster log parsing. Service dependency graph: Visualize relationships between services, detect bottlenecks, and optimize network flows. These dashboards provide filtering options to isolate specific logs, such as DNS errors or traffic patterns, enabling efficient root cause analysis. Summary statistics and top-level metrics further enhance understanding of cluster health and activity. Fig 2: Azure managed Grafana dashboard for container network logs Conclusion Container network logs for AKS offer a powerful and cost optimized way to monitor and analyze network activity, enhance troubleshooting, security, and ensure compliance. To get started, enable Advanced Container Networking Services in your AKS cluster and configure custom resources for logging. Visualize your logs in Grafana dashboards and Azure Log Analytics to unlock actionable insights. Learn more here.Azure CNI now supports Node Subnet IPAM mode with Cilium Dataplane
Azure CNI Powered by Cilium is a high-performance data plane leveraging extended Berkeley Packet Filter (eBPF) technologies to enable features such as network policy enforcement, deep observability, and improved service routing. Legacy CNI supports Node Subnet where every pod gets an IP address from a given subnet. AKS clusters that require VNet IP addressing mode (non-overlay scenarios) are typically advised to use Pod Subnet mode. However, AKS clusters that do not face the risk of IP exhaustion can continue to use node subnet mode for legacy reasons and switch the CNI dataplane to utilize Cilium's features. With this feature launch, we are providing that migration path! Users often leverage node subnet mode in Azure Kubernetes Service (AKS) clusters for ease of use. This mode provides an optionality where users do not want to worry about managing multiple subnets, especially when using smaller clusters. Besides, let’s highlight some additional benefits unlocked by this feature. Improved Network Debugging Capabilities through Advanced Container Networking Services By upgrading to Azure CNI Powered by Cilium with Node Subnet, Advanced Container Networking Services opens the possibility of using eBPF tools to gather request metrics at the node and pod level. Advanced Observability tools provide a managed Grafana dashboard to inspect these metrics for a streamlined incident response experience. Advanced Network Policies Network policies for Legacy CNI present a challenge because policies on IP-based filtering require constant updating in a Kubernetes cluster where pod IP addresses frequently change. Enabling the Cilium data plane offers an efficient and scalable approach to managing network policies. Create an Azure CNI Powered by Cilium cluster with node subnet as the IP Address Management (IPAM) networking model. This is the default option when done with a `--network-plugin azure` flag. az aks create --name <clusterName> --resource-group <resourceGroupName> --location <location> --network-plugin azure --network-dataplane cilium --generate-ssh-keys A flat network can lead to less efficient use of IP addresses. Careful planning through the List Usage command of a given VNet helps to see current usage of the subnet space. AKS creates a VNet and subnet automatically from cluster creation. Note the resource group for this VNet is generated based on the resource group for the cluster, the cluster name, and location. From the Portal under Settings > Networking for the AKS cluster, we can see the names of the resources created automatically. az rest --method get \ --url https://management.azure.com/subscriptions/{subscription-id} /resourceGroups/MC_acn-pm_node-subnet-test_westus2/providers/Microsoft.Network/virtualNetworks/aks-vnet-34761072/usages?api-version=2024-05-01 { "value": [ { "currentValue": 87, "id": "/subscriptions/9b8218f9-902a-4d20-a65c-e98acec5362f/resourceGroups/MC_acn-pm_node-subnet-test_westus2/providers/Microsoft.Network/virtualNetworks/aks-vnet-34761072/subnets/aks-subnet", "isAdjustable": false, "limit": 65531, "name": { "localizedValue": "Subnet size and usage", "value": "SubnetSpace" }, "unit": "Count" } ] } To better understand this utilization, click the link of the Virtual network then access the list of Connected Devices. The view also shows which IPs are utilized on a given node. There are a total of 87 devices consistent with the previous command line output of subnet usage. Since the default creates three nodes with a max pod count of 30 per node (configurable up to 250), IP exhaustion is not a concern although careful planning is required for larger clusters. Next, we will enable Advanced Container Networking Services (ACNS) on this cluster. az aks update --resource-group <resourceGroupName> --name <clusterName> --enable-acns Create a default deny Cilium Network policy. The namespace is `default`, and we will use `app: server` as the label in this example. kubectl apply -f - <<EOF apiVersion: cilium.io/v2 kind: CiliumNetworkPolicy metadata: name: default-deny namespace: default spec: endpointSelector: matchLabels: app: server ingress: - {} egress: - {} EOF The empty brackets under ingress and egress represent all traffic. Next, we will use `agnhost`, a network connectivity utility used in Kubernetes upstream testing that can help set up a client/server scenario. kubectl run server --image=k8s.gcr.io/e2e-test-images/agnhost:2.41 --labels="app=server" --port=80 --command -- /agnhost serve-hostname --tcp --http=false --port "80" Get the server address IP: kubectl get pod server -o wide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES server 1/1 Running 0 9m 10.224.0.57 aks-nodepool1-20832547-vmss000002 <none> <none> Create a client that will use the agnhost utility to test the network policy. Open a new terminal window as this will also open a new shell. kubectl run -it client --image=k8s.gcr.io/e2e-test-images/agnhost:2.41 --command -- bash Test connectivity to the server from client. Timeout is expected since the network policy is default deny for all traffic in the default namespace. Your pod IP may be different from the example. bash-5.0# ./agnhost connect 10.224.0.57:80 --timeout=3s --protocol=tcp –verbose TIMEOUT Remove the network policy. In practice, additional policies would be added to retain the default deny policy while allowing applications that satisfy the conditions to allow connectivity. kubectl delete cnp default-deny From a shell with the client pod, verify the connection is now allowed. If successful, there is simply no output. kubectl attach client -c client -i -t bash-5.0# ./agnhost connect 10.224.0.57:80 --timeout=3s --protocol=tcp Connectivity between server and client is restored. Additional tools such as Hubble UI for debugging can be found in Container Network Observability - Advanced Container Networking Services (ACNS) for Azure Kubernetes Service (AKS) - Azure Kubernetes Service | Microsoft Learn. Conclusion Building a seamless migration path is critical to continued growth and adoption of ACPC. The goal is to provide a best-in-class experience by providing an upgrade path to enable the Cilium data plane to enable high-performance networking across various IP addressing modes. This allows for flexibility to fit your IP address plans to build varied workload types using AKS networking. Keep an eye out on the AKS public roadmap for more developments in the near future. Resources Learn more about Azure CNI Powered by Cilium. Learn more about IP address planning. Visit Azure CNI Powered by Cilium benchmarking to see performance benchmarks using an eBPF dataplane. Visit to learn more about Advanced Container Networking Services.What's New in the World of eBPF from Azure Container Networking!
Azure Container Networking Interface (CNI) continues to evolve, now bolstered by the innovative capabilities of Cilium. Azure CNI Powered by Cilium (ACPC) leverages Cilium’s extended Berkeley Packet Filter (eBPF) technologies to enable features such as network policy enforcement, deep observability, and improved service routing. Here’s a deeper look into the latest features that make management of Azure Kubernetes Service (AKS) clusters more efficient, scalable, and secure. Improved Performance: Cilium Endpoint Slices One of the standout features in the recent updates is the introduction of CiliumEndpointSlice. This feature significantly enhances the performance and scalability of the Cilium dataplane in AKS clusters. Previously, Cilium used Custom Resource Definitions (CRDs) called CiliumEndpoints to manage pods. Each pod had a CiliumEndpoint associated with it, which contained information about the pod’s status and properties. However, this approach placed significant stress on the control plane, especially in larger clusters. To alleviate this load, CiliumEndpointSlice batches CiliumEndpoints and their updates, reducing the number of updates propagated to the control plane. Our performance testing has shown remarkable improvements: Average API Server Responsiveness: Upto 50% decrease in latency, meaning faster processing of queries. Pod Startup Latencies: Upto 60% reduction, allowing for faster deployment and scaling. In-Cluster Network Latency: Upto 80% decrease, translating to better application performance. Note that this feature is Generally Available in AKS clusters, by default, using Cilium 1.17 release and above and does not require additional configuration changes! Learn more about improvements unlocked by CiliumEndpointSlices with Azure CNI by Cilium - High-Scale Kubernetes Networking with Azure CNI Powered by Cilium | Microsoft Community Hub. Deployment Flexibility: Dual Stack for Cilium Network Policies Kubernetes clusters operating on an IPv4/IPv6 dual-stack network enable workloads to natively access both IPv4 and IPv6 endpoints without incurring additional complexities or performance drawbacks. Previously, we had enabled the use of dual stack networking on AKS clusters (starting with AKS 1.29) running Azure CNI powered by Cilium in preview mode. Now, we are happy to announce that the feature is Generally Available! By enabling both IPv4 and IPv6 addressing, you can manage your production AKS clusters in mixed environments, accommodating various network configurations seamlessly. More importantly, dual-stack support in Azure CNI’s Cilium network policies extend security benefits for AKS clusters in those complex environments. For instance, you can enable dual stack AKS clusters using eBPF dataplane as follows: az aks create \ --location <region> \ --resource-group <resourceGroupName> \ --name <clusterName> \ --network-plugin azure \ --network-plugin-mode overlay \ --network-dataplane cilium \ --ip-families ipv4,ipv6 \ --generate-ssh-keys Learn more about Azure CNI’s Network Policies - Configure Azure CNI Powered by Cilium in Azure Kubernetes Service (AKS) - Azure Kubernetes Service | Microsoft Learn Ease of Use: Node Subnet Mode with Cilium Azure CNI now supports Node Subnet IPAM mode with Cilium Dataplane. In Node Subnet mode IP addresses to pods are assigned from the same subnet as the node itself, simplifying routing and policy management. This mode is particularly beneficial for smaller clusters where managing multiple subnets is cumbersome. AKS clusters using this mode also gain the benefits of improved network observability, Cilium Network Policies and FQDN filtering and more capabilities unlocked by Advanced Container Networking Services (ACNS). More notable, with this feature we now support all IPAM configuration options with eBPF dataplane on AKS clusters. You can create an AKS cluster with node subnet IPAM mode and eBPF dataplane as follows: az aks create \ --name <clusterName> \ --resource-group <resourceGroupName> \ --location <location> \ --network-plugin azure \ --network-dataplane cilium \ --generate-ssh-keys Learn more about Node Subnet - Configure Azure CNI Powered by Cilium in Azure Kubernetes Service (AKS) - Azure Kubernetes Service | Microsoft Learn. Defense-in-depth: Cilium Layer 7 Policies Azure CNI by Cilium extends its comprehensive Layer4 network policy capabilities to Layer7, offering granular control over application traffic. This feature enables users to define security policies based on application-level protocols and metadata, adding a powerful layer of security and compliance management. Layer7 policies are implemented using Envoy, an open-source service proxy, which is part of ACNS Security Agent operating in conjunction with the Cilium agent. Envoy handles traffic between services and provides necessary visibility and control at the application layer. Policies can be enforced based on HTTP and gRPC methods, paths, headers, and other application-specific attributes. Additionally, Cilium Network Policies support Kafka-based workflows, enhancing security and traffic management. This feature is currently in public preview mode and you can learn more about the getting started experience here - Introducing Layer 7 Network Policies with Advanced Container Networking Services for AKS Clusters! | Microsoft Community Hub. Coming Soon: Transparent Encryption with Wireguard By leveraging Cilium’s Wireguard, customers can achieve regulatory compliance by ensuring that all network traffic, whether HTTP-based or non-HTTP, is encrypted. Users can enable inter-node transparent encryption in their Kubernetes environments using Cilium’s open-source based solution. When Wireguard is enabled, the cilium agent on each cluster node will establish a secure Wireguard tunnel with all other known nodes in the cluster to encrypt traffic between cilium endpoints. This feature will soon be in public preview and will be enabled as part of ACNS. Stay tuned for more details on this. Conclusion These new features in Azure CNI Powered by Cilium underscore our commitment to enhancing default network performance and security in your AKS environments, all while collaborating with the open-source community. From the impressive performance boost with CiliumEndpointSlice to the adaptability of dual-stack support and the advanced security of Layer7 policies and Wireguard based encryption, these innovations ensure your AKS clusters are not just ready for today but are primed for the future. Also, don’t forget to dive into the fascinating world of eBPF-based observability in multi-cloud environments! Check out our latest post - Retina: Bridging Kubernetes Observability and eBPF Across the Clouds. Why wait, try these out now! Stay tuned to the AKS public roadmap for more exciting developments! For additional information, visit the following resources: For more info about Azure CNI Powered by Cilium visit - Configure Azure CNI Powered by Cilium in AKS. For more info about ACNS visit Advanced Container Networking Services (ACNS) for AKS | Microsoft Learn.
Introducing Layer 7 Network Policies with Advanced Container Networking Services for AKS Clusters!
We have been on an exciting journey to enhance the network observability and security capabilities of Azure Kubernetes Service (AKS) clusters through our Azure Container Networking Services offering. Our initial step, the launch of Fully Qualified Domain Name (FQDN) filtering, marked a foundational step in enabling policy-driven egress control. By allowing traffic management at the domain level, we set the stage for more advanced and fine-grained security capabilities that align with modern, distributed workloads. This was just the beginning, a glimpse into our commitment to enabling AKS users with robust and granular security controls. Today, we are thrilled to announce the public preview of Layer 7 (L7) Network Policies for AKS and Azure CNI powered by Cilium users with Advanced Container Networking Services enabled. This update brings a whole new dimension of security to your containerized environments, offering much more granular control over your application layer traffic. Overview of L7 Policy Unlike traditional Layer 3 and Layer 4 policies that operate at the network and transport layers, L7 policies operate at the application layer. This enables more precise and effective traffic management based on application-specific attributes. L7 policies enable you to define security rules based on application-layer protocols such as HTTP(S), gRPC, and Kafka. For example, you can create policies that allow traffic based on HTTP(S) methods (GET, POST, etc.), headers, paths, and other protocol-specific attributes. This level of control helps in implementing fine-grained access control, restricting traffic based on the actual content of the communication, and gaining deeper insights into your network traffic at the application layer. Use cases of L7 policy API Security: For applications exposing APIs, L7 policies provide fine-grained control over API access. You can define policies that only allow specific HTTP(S) methods (e.g., GET for read-only operations, POST for creating resources) on particular API paths. This helps in enforcing API security best practices and prevent unnecessary access. Zero-Trust Implementation: L7 policies are a key component in implementing a Zero-Trust security model within your AKS environment. By default-denying all traffic and then explicitly allowing only necessary communication based on application-layer context, you can significantly reduce the attack surface and improve overall security posture. Microservice Segmentation and Isolation: In microservice architecture, it's essential to isolate services to limit the blast radius of potential security breaches. L7 policies allow you to define precise rules for inter-service communication. For example, you can ensure that a billing service can only be accessed by an order processing service via specific API endpoints and HTTP(S) methods, preventing unauthorized access from other services. How Does It Work? When a pod sends out network traffic, it is first checked against your defined rules using a small, efficient program called an extended Berkley Packet Filter (eBPF) probe. This probe marks the traffic if L7 policies are enabled for that pod, it is then redirected to a local Envoy Proxy. The Envoy proxy here is part of the ACNS Security Agent, separate from the Cilium agent. The Envoy Proxy, part of the ACNS Security Agent, then acts as a gatekeeper, deciding whether the traffic is allowed to proceed based on your policy criteria. If the traffic is permitted, it flows to its destination. If not, the application receives an error message saying access denied. Example: Restricting HTTP POST Requests Let's say you have a web application running on your AKS cluster, and you want to ensure that a specific backend service (backend-app) only accepts GET requests on the /data path from your frontend application (frontend-app). With L7 Network Policies, you can define a policy like this: apiVersion: cilium.io/v2 kind: CiliumNetworkPolicy metadata: name: allow-get-products spec: endpointSelector: matchLabels: app: backend-app # Replace with your backend app name ingress: - fromEndpoints: - matchLabels: app: frontend-app # Replace with your frontend app name toPorts: - ports: - port: "80” protocol: TCP rules: http: - method: "GET" path: "/data" How Can You Observe L7 Traffic? Advanced Container Networking Services with L7 policies also provides observability into L7 traffic, including HTTP(S), gRPC, and Kafka, through Hubble. Hubble is enabled by default with Advanced Container Networking Services. To facilitate easy analysis, Advanced Container Networking Services with L7 policy offers pre-configured Azure Managed Grafana dashboards, located in the Dashboards > Azure Managed Prometheus folder. These dashboards, such as "Kubernetes/Networking/L7 (Namespace)" and "Kubernetes/Networking/L7 (Workload)", provide granular visibility into L7 flow data at the cluster, namespace, and workload levels. The screenshot below shows a Grafana dashboard visualizing incoming HTTP traffic for the http-server-866b29bc75 workload in AKS over the last 5 minutes. It displays request rates, policy verdicts (forwarded/dropped), method/status breakdown, and dropped flow rates for real-time monitoring and troubleshooting. This is just an example; similar detailed metrics and visualizations, including heatmaps, are also available for gRPC and Kafka traffic on our pre-created dashboards. For real-time log analysis, you can also leverage the Hubble CLI and UI options offered as part of the Advanced Container Networking Services observability solution, allowing you to inspect individual L7 flows and troubleshoot any policy enforcement. Call to Action We encourage you to try out the public preview of L7 Network Policies on your AKS clusters and level up your network security controls for containerized workloads. We value your feedback as we continue to develop and improve this feature. Please refer to the Layer 7 Policy Overview for more information and visit How to Apply L7 Policy for an example scenario.
High-Scale Kubernetes Networking with Azure CNI Powered by Cilium
Kubernetes users have diverse cluster networking needs, but paramount to them all are efficient pod networking and robust security features. Azure CNI (Container Networking Interface) powered by Cilium is a solution that addresses these needs by blending the capabilities of the Azure CNI’s control plane and Cilium’s eBPF dataplane. Cilium enables performant networking and security by leveraging the power of eBPF (extended Berkeley Packet Filter), a revolutionary Linux kernel technology. eBPF enables the execution of custom code within the Linux kernel, providing both flexibility and efficiency. This translates to: High-performance networking: eBPF enables efficient packet processing, reduced latencies and improved throughput Enhanced security: Azure CNI (AzCNI) powered by Cilium enables network security policies based on DNS to easily manage and secure network traffic through Advanced Network Security features. Better observability: Our eBPF-based Advanced Network Observability suite provides detailed monitoring, tracing and diagnostics tools for cluster users. Introducing CiliumEndpointSlice A performant CNI dataplane is crucial for low-latency, high-throughput pod communication, enhancing distributed application efficiency and user experience. While Cilium’s eBPF-powered dataplane provides high-performance networking today, we sought to further enhance its scalability and performance. To do this, we looked at enabling a new feature in the dataplane’s configuration, CiliumEndpointSlice, thereby achieving: Lower traffic load on the Kubernetes control plane, leading to reduced control plane memory consumption and improved performance Faster pod start-up latencies Faster in-cluster network latencies for better application performance In particular, this feature improves upon how Azure CNI powered by Cilium manages pods. Previously, Cilium managed pods using Custom Resource Definitions (CRDs) called CiliumEndpoints. Each pod has a CiliumEndpoint associated with it. The CRD contains information about a pod’s status and properties. The Cilium Agent, a core component of the dataplane, runs on every node and watches each of these CiliumEndpoints for information about updates to pods. We have observed that this behavior can place significant stress and load on the control plane, leading to performance bottlenecks, especially for larger clusters. To alleviate load on the control plane, we are bringing in CiliumEndpointSlice, a feature which batches CiliumEndpoints and their associated updates. This reduces the number of updates propagated to the control plane. Consequently, we greatly reduce the risk of overloading the control plane at scale, ensuring smoother operation of the cluster. Performance Testing We have conducted performance testing of Azure CNI powered by Cilium with and without CiliumEndpointSlice enabled. The testing was done on a cluster with the following dimensions: 1000 nodes (Standard_D4_v3) 20,000 pods (i.e. 20 pods per node) 1 service with 4000 backends 800 services with 20 backends each The test involved repeating the following actions 10 times: creating deployments and services, restarting deployments, and deleting deployments and services. We detail the various performance metrics measured below. Average APIServer Responsiveness This metric measures the average latency of responses to LIST requests (one of the most expensive types of requests to the control plane) by the kube-apiserver. With CiliumEndpointSlice enabled, we observed a remarkable 50% decrease in latency, dropping from an average of ~1.5 seconds to ~0.25 seconds! For cluster users, this means much faster processing of queries sent to the kube-apiserver, leading to improved performance. Pod Startup Latencies This metric measures the time taken for a pod to be reported as running. Here, an over 60% decrease in pod startup latency was observed with CiliumEndpointSlice enabled, allowing for faster deployment and scaling of applications. In-Cluster Network Latency This is a critical metric, measuring the latency of pings from a prober pod to a server. An over 80% decrease in latency was observed. This reduction in latency translates to better application performance. Azure CNI powered by Cilium offers a powerful eBPF-based solution for Kubernetes networking and security. With the enablement of CiliumEndpointSlice from Kubernetes version 1.32 on AzCNI Powered by Cilium clusters, we see further improvements in application and control plane performance. For more information, visit https://learn.microsoft.com/en-us/azure/aks/azure-cni-powered-by-cilium.eBPF-Powered Observability Beyond Azure: A Multi-Cloud Perspective with Retina
Kubernetes simplifies container orchestration but introduces observability challenges due to dynamic pod lifecycles and complex inter-service communication. eBPF technology addresses these issues by providing deep system insights and efficient monitoring. The open-source Retina project leverages eBPF for comprehensive, cloud-agnostic network observability across AKS, GKE, and EKS, enhancing troubleshooting and optimization through real-world demo scenarios.Securing Microservices with Cilium and Istio
The adoption of Kubernetes and containerized applications is booming, leading to new challenges in visibility and security. As the landscape of cloud-native applications is rapidly evolving so are the number of sophisticated attacks targeting containerized workloads. Traditional tools often fall short in tracking the usage and traffic flows within these applications. The immutable nature of container images and the short lifespan of containers further necessitate addressing vulnerabilities early in the delivery pipeline. Comprehensive Security Controls in Kubernetes Microsoft Azure offers a range of security controls to ensure comprehensive protection across various layers of the Kubernetes environment. These controls include but are not limited to: Cluster Security: Features such as private clusters, managed cluster identity, and API server authorized ranges enhance security at the cluster level. Node and Pod Security: Hardened bootstrapping, confidential nodes, and pod sandboxing are implemented to secure the nodes and pods within a cluster. Network Security: Advanced Container Networking Services and Cilium Network policies offer granular control over network traffic. Authentication and Authorization: Azure Policy in-cluster enforcement, Entra authentication, and Istio mTLS and authorization policies provide robust identity and access management. Image Scanning: Microsoft Defender for Cloud provides both image and runtime scanning to identify vulnerabilities and threats. Let’s highlight how you can secure micro services while scaling your applications running on Azure Kubernetes Service (AKS) using service mesh for robust traffic management, and network policies for security. Micro segmentation with Network Policies Micro segmentation is crucial for enhancing security within Kubernetes clusters, allowing for the isolation of workloads and controlled traffic between microservices. Azure CNI by Cilium leverages eBPF to provide high-performance networking, security, and observability features. It dynamically inserts eBPF bytecode into the Linux kernel, offering efficient and flexible control over network traffic. Cilium Network Policies enable network isolation within and across Kubernetes clusters. Cilium also provides an identity-based security model, offering Layer 7 (L7) traffic control, and integrates deep observability for L4 to L7 metrics in Kubernetes clusters. A significant advantage of using Azure CNI based on Cilium is its seamless integration with existing AKS environments, requiring minimal modifications to your infrastructure. Note that Cilium Clusterwide Network Policy (CCNP) is not supported at the time of writing this blog post. FQDN Filtering with Advanced Container Networking Services (ACNS) Traditional IP-based policies can be cumbersome to maintain. ACNS allows for DNS-based policies, providing a more granular and user-friendly approach to managing network traffic. This is supported only with Azure CNI powered by Cilium and includes a security agent DNS proxy for FQDN resolution even during upgrades. It’s worth noting that with Cilium’s L7 enforcement, you can control traffic based on HTTP methods, paths, and headers, making it ideal for APIs, microservices, and services that use protocols like HTTP, gRPC, or Kafka. At the time of writing this blog, this capability is not supported via ACNS. More on this in a future blog! AKS Istio Add-On: Mutual TLS (mTLS) and Authorization Policy Istio enhances the security of microservices through its built-in features, including mutual TLS (mTLS) and authorization policies. The Istiod control plane, acting as a certificate authority, issues X.509 certificates to the Envoy sidecar proxies via the Secret Discovery Service (SDS). Integration with Azure Key Vault allows for secure management of root and intermediate certificates. The PeerAuthentication Custom Resource in Istio controls the traffic accepted by workloads. By default, it is set to PERMISSIVE to facilitate migration but can be set to STRICT to enforce mTLS across the mesh. Istio also supports granular authorization policies, allowing for control over IP blocks, namespaces, service accounts, request paths, methods, and headers. The Istio add-on also supports integration with Azure Key Vault (AKV) and the AKV Secrets Store CSI Driver Add-On for plug-in CA certificates, where the root CA lives on a secure machine offline, and the intermediate certs for the Istiod control plane are synced to the cluster by the CSI Driver Add-On. Additionally, certificates for the Istio ingress gateway for TLS termination or SNI passthrough can also be stored in AKV. Defense-In-Depth with Cilium, ACNS and Istio Combining the capabilities of Cilium's eBPF technologies through ACNS and AKS managed Istio addon, AKS provides a defense-in-depth strategy for securing Kubernetes clusters. Azure CNI's Cilium Network Policies and ACNS FQDN filtering enforce Pod-to-Pod and Pod-to-egress policies at Layer 3 and 4, while Istio enforces STRICT mTLS and Layer 7 authorization policies. This multi-layered approach ensures comprehensive security coverage across all layers of the stack. Now, let’s highlight the key steps in achieving this: Step 1: Create an AKS Cluster with Azure CNI (by Cilium), ACNS and Istio Addon enabled. az aks create \ --resource-group $RESOURCE_GROUP \ --name $CLUSTER_NAME \ --location $LOCATION \ --kubernetes-version 1.30.0 \ --node-count 3 \ --node-vm-size standard_d16_v3 \ --enable-managed-identity \ --network-plugin azure \ --network-dataplane cilium \ --network-plugin-mode overlay \ --pod-cidr 192.168.0.0/16 \ --enable-asm \ --enable-acns \ --generate-ssh-keys Step 2: Create Cilium FQDN policy that allows egress traffic to google.com while blocking traffic to httpbin.org. Sample Policy (fqdn-filtering-policy.yaml): apiVersion: cilium.io/v2 kind: CiliumNetworkPolicy metadata: name: sleep-network-policy namespace: foo spec: endpointSelector: matchLabels: app: sleep egress: - toFQDNs: - matchPattern: "*.google.com" - toEndpoints: - matchLabels: "k8s:io.kubernetes.pod.namespace": foo "k8s:app": helloworld - toEndpoints: - matchLabels: "k8s:io.kubernetes.pod.namespace": kube-system "k8s:k8s-app": kube-dns toPorts: - ports: - port: "53" protocol: ANY Apply policy: kubectl apply -f fqdn-filtering-policy.yaml Step 3: Create an Istio deny-by-default AuthorizationPolicy. This denies all requests across the mesh unless specifically authorized with an “ALLOW” policy. Sample Policy (istio-deny-all-authz.yaml): apiVersion: security.istio.io/v1 kind: AuthorizationPolicy metadata: name: allow-nothing namespace: aks-istio-system spec: {} Apply policy: kubectl apply -f istio-deny-all-authz.yaml Step 4: Deploy an Istio L7 AuthorizationPolicy to explicitly allow traffic to the “sample” pod in namespace foo for http “GET” requests. Sample Policy (istio-L7-allow-policy.yaml): apiVersion: security.istio.io/v1 kind: AuthorizationPolicy metadata: name: allow-get-requests namespace: foo spec: selector: matchLabels: app: sample action: ALLOW rules: - to: - operation: methods: [“GET”] Apply policy: kubectl apply -f istio-L7-allow-policy.yaml Step 5: Deploy an Istio strict mTLS PeerAuthentication Resource to enforce that all workloads in the mesh only accept Istio mTLS traffic. Sample PeerAuthentication (istio-peerauth.yaml): apiVersion: security.istio.io/v1 kind: PeerAuthentication metadata: name: strict-mtls namespace: aks-istio-system spec: mtls: mode: STRICT Apply policy: kubectl apply -f istio-peerauth.yaml These examples demonstrate how you can manage traffic to specific FQDNs and enforce L7 authorization rules in your AKS cluster. Conclusion Traditional IP and perimeter security models are insufficient for the dynamic nature of cloud-native environments. More sophisticated security mechanisms, such as identity-based policies and DNS names, are required. Azure CNI, powered by Cilium and ACNS, provides robust FQDN filtering and Layer 3/4 Network Policy enforcement. The Istio add-on offers mTLS for identity-based encryption and Layer7 authorization policies. A defense-in-depth model, incorporating both Azure CNI and service mesh mechanisms, is recommended for maximizing security posture. So, give these a try and let us know (Azure Kubernetes Service Roadmap (Public)) how we can evolve our roadmap to help you build the best with Azure. Credit(s): Niranjan Shankar, Sr. Software Engineer, MicrosoftSecure, High-Performance Networking for Data-Intensive Kubernetes Workloads
In today’s data-driven world, AI and high-performance computing (HPC) workloads demand a robust, scalable, and secure networking infrastructure. As organizations rely on Kubernetes to manage these complex workloads, the need for advanced network performance becomes paramount. In this blog series, we explore how Azure CNI powered by Cilium, built on eBPF technology, is transforming Kubernetes networking. From high throughput and low latency to enhanced security and real-time observability, discover how these cutting-edge advancements are paving the way for secure, high-performance AI workloads. Ready to optimize your Kubernetes clusters?
Enhancing Observability with Inspektor Gadget
Thorough observability is essential to a pain free cloud experience. Azure provides many general-purpose observability tools, but you may want to create custom tooling . Inspektor Gadget is an open-source framework that makes customizable data collection easy. Microsoft recently contributed new features to Inspektor Gadget that further enhance its modular framework, making it even easier to meet your specific systems inspection needs. Of course, we also made it easy for Azure Kubernetes Service (AKS) users to use.Use cases of Advanced Network Observability for your Azure Kubernetes Service clusters
The blog explores the use cases of Advanced Network Observability for Azure Kubernetes Service (AKS) clusters. It introduces the Advanced Network Observability feature, which brings Hubble's control plane to both Cilium and Non-Cilium Linux data planes. This feature provides deep insights into containerized workloads, enabling precise detection and root-cause analysis of network-related issues in Kubernetes clusters. The document also includes customer scenarios that demonstrate the benefits of Advanced Network Observability, such as DNS metrics, network policy drops at the pod level, and traffic imbalance for pods within a workload
