Kubernetes Overview
Tenable offers various methods for connecting your Kubernetes environment. The best method for your organization depends on your cluster architecture, security requirements, and the specific features you want to enable, such as the Kubernetes Admission Controller or sensor-based scanning.
- Cloud-based (Agentless): This method uses cloud-native APIs (such as Amazon EKS, Azure AKS, Google GKE, or Oracle OKE) to access the Kubernetes API through cloud IAM. Use this method to discover and scan clusters without installing software in your environment.
- Helm chart-based (Sensor-based): This method is required for private or on-premises clusters. It is also required to enable the Admission Controller and other Kubernetes Security Features.
For more details, see Kubernetes.
Cluster Discovery
Tenable discovers and displays clusters in the Inventory page through three distinct methods based on the hosting environment:
- Cloud-managed clusters (AWS, Azure, GCP, and OCI): Tenable automatically discovers these clusters and displays them in the Inventory page as soon as you connect the associated cloud account.
- Unsupported cloud providers — If you use a cloud provider that the platform does not currently support for API-based discovery, you can connect the cluster using a Helm chart. Clusters connected this way are associated with an account classified as On-Premises within the Console.
- Managed (Private) and on-premises clusters: These environments appear in the Inventory page only after you connect the cluster to Tenable via Helm chart.
Cluster Lifecycle
How a cluster that is stale is deleted from the Tenable One Cloud Exposure Console depends on how you connected it.
Discovered with Agentless
Clusters discovered through a cloud account that uses cloud-based agentless connection (EKS, AKS, GKE, and OKE) are removed automatically when they are deleted in your cloud environment. A failed sync does not remove a cluster.
Discovered with Agents Only
Clusters that only have a sensor deployed via Helm chart have no cloud account to confirm they still exist. Tenable relies on the cluster reporting in. Based on its reporting activity, a cluster moves through the following lifecycle stages. The Console does not display a stage name. To see where a cluster is in its lifecycle, check whether it appears in Inventory and review its Cluster Sync Time column.
| Lifecycle Stage | What You See in the Console | What Happens |
|---|---|---|
| Reporting | The cluster appears in your inventory and its findings are open. Cluster Sync Time is within the last 24 hours. | The cluster has reported within the last 24 hours. |
| Not reporting | The cluster still appears in your Inventory and its findings stay open. Cluster Sync Time is older than 24 hours. | The cluster has not reported for more than 24 hours. The countdown to deletion from the retention period starts. See Retention Period Settings. |
| Soft deleted | The cluster no longer appears in your Inventory and its findings are closed. | The cluster did not report for the full retention period (30 days by default), so Tenable deletes it from the Console. |
| Restored | The cluster is back in Inventory with its configuration and history, and its findings are reopened. Cluster Sync Time is within the last 24 hours. | The cluster reports again before it is permanently deleted. |
| Permanently deleted | The cluster is permanently deleted from the Console, its findings stay closed, and it can not be restored. | 30 days after the cluster is deleted from the Console, it is permanently deleted and cannot be restored. |
Retention Period Settings
In Settings > Workload > Kubernetes Cluster tab, you can enable or disable cluster retention and set the retention period. You can also override the retention period for individual scopes, for example to remove short-lived test clusters sooner or to keep clusters that go offline for extended maintenance. See Configure Self-Managed Kubernetes Cluster Retention.
Kubernetes Security Features
While discovery provides high-level visibility, you can enable additional security features to better protect your environment:
- Resource Sync — Provides visibility into your Kubernetes resources by synchronizing cluster data with Tenable. This ensures your resource inventory reflects the current state of your environment.
- Runtime Detection — Monitors system-level activity via an eBPF sensor to detect malicious or suspicious activity in near-real time.
- Admission Controller — Intercepts API requests to ensure only compliant resources are deployed based on your security policies.
- Workload Protection — Scans Kubernetes nodes and running container images for vulnerabilities and security misconfigurations.
Use the following sections to determine whether to use the Helm chart (sensor-based) or cloud-native (agentless) methods for resource synchronization, runtime detection, workload protection, and additional security features.
When Is Helm Chart Integration Required?
A Helm chart integration moves your security posture beyond basic metadata discovery to a comprehensive, deep-scan analysis. You should choose this integration method in the following scenarios:
- Establish connectivity — Use this method if your cluster is:
- Managed (Private) — You use a managed (private) cluster where the API server is not accessible from the internet.
- On-premises — You host your infrastructure on customer-managed resources, such as private data centers.
- Enable Kubernetes-specific features — Activate features such as Runtime Detection, Admission Controller, and Vulnerability Management.
-
Meet advanced security or compliance requirements, such as:
- Real-time policy enforcement — Required for the Kubernetes Admission Controller.
- Deep configuration visibility — Provides granular insights into Role-Based Access Control (RBAC) and cluster configurations.
For more details, see Connect Kubernetes Clusters.
Connect Kubernetes Clusters
Before you connect your cluster, determine if your deployment model requires the Helm chart method. Use the Kubernetes cluster Helm chart for private clusters, on-premises environments, and specific security features where cloud-native API methods are not supported.
Kubernetes Resource Sync and Workload Protection
| Cloud Provider | Managed (Public) | Managed (Private) | Unmanaged |
|---|---|---|---|
| AWS 1 | Grant relevant permissions (agentless) | Install via Helm chart (sensor) | Install via Helm chart (sensor) |
| Azure 2 | Grant relevant permissions (agentless) | Install via Helm chart (sensor) | Install via Helm chart (sensor) |
| GCP 3 | Grant relevant permissions (agentless) | Install via Helm chart (sensor) | Install via Helm chart (sensor) |
| OCI | Grant relevant permissions (agentless) | Install via Helm chart (sensor) | Install via Helm chart (sensor) |
| On-Premises and any unsupported cloud platform | N/A | N/A | Install via Helm chart (sensor) |
The following features are currently not supported for vulnerability management scanning:
AWS 1 - Fargate is not supported for vulnerability management.
Azure 2 - Virtual nodes are not supported for vulnerability management.
GCP 3 - Autopilot is not supported for vulnerability management.
Runtime Detection
To use Runtime Detection, you must install the sensor by using Helm. This requirement applies to all supported cloud providers across public, private, or unmanaged Kubernetes and OpenShift clusters.
If you are already using the Helm-based sensor for connectivity, you do not need to perform a separate installation. Instead, update your existing deployment with the runtime flag. When you perform the installation or update, include the --set sensor.runtimeDetection=true flag to enable runtime security.
For more details, see Install the Runtime Detection Sensor for Kubernetes.
Admission Controller
To use the Admission Controller you must install the sensor by using Helm. This requirement applies to all supported cloud providers across public, private, or unmanaged Kubernetes and OpenShift clusters.
For more details, see Install Admission Controller.