惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

S
SegmentFault 最新的问题
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
B
Blog RSS Feed
Y
Y Combinator Blog
T
Tailwind CSS Blog
博客园 - 三生石上(FineUI控件)
J
Java Code Geeks
Stack Overflow Blog
Stack Overflow Blog
aimingoo的专栏
aimingoo的专栏
Jina AI
Jina AI
The GitHub Blog
The GitHub Blog
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
A
About on SuperTechFans
H
Hackread – Cybersecurity News, Data Breaches, AI and More
D
Docker
酷 壳 – CoolShell
酷 壳 – CoolShell
C
Check Point Blog
M
MIT News - Artificial intelligence
Last Week in AI
Last Week in AI
V
V2EX
腾讯CDC
F
Fortinet All Blogs
博客园 - 叶小钗
T
The Blog of Author Tim Ferriss

Datadog | The Monitor blog

Introducing our open source AI-native SAST Instrument and monitor Boomi integration flows with OpenTelemetry and Datadog Not all index scans are equal: How we cut query latency by over 99% Platform engineering metrics: What to measure and what to ignore Integrate Recorded Future threat intelligence with Datadog Cloud SIEM CI/CD security: threat modeling using a MITRE-style threat matrix CI/CD security: How to secure your GitHub ecosystem Ingress NGINX is EOL: A practical guide for migrating to Kubernetes Gateway API Operating agentic AI with Amazon Bedrock AgentCore and Datadog LLM Observability: Lessons from NTT DATA Introducing the Datadog Code Security MCP Capture and analyze custom heatmaps in Session Replay Understand session replays faster with AI summaries and smart chapters Monitor ClickHouse query performance with Datadog Database Monitoring How we designed empathetic alert sounds for on-call engineers Search and act across Datadog to resolve issues faster with Bits Assistant Measure the business impact of every product change with Datadog Experiments Analyzing round trip query latency Configuring JavaScript caches for better performance Introducing Bits AI Dev Agent for Code Security Datadog achieves ISO 42001 certification for responsible AI Monitor Nutanix clusters, hosts, and VMs with Datadog Monitor Juniper Mist in Datadog A new Host Map for modern infrastructure Annotate traces to improve LLM quality with Datadog LLM Observability What’s new in Cloud SIEM: AI-powered investigations, enhanced threat intelligence, and scalable security operations Explore Kubernetes with native OpenTelemetry data Monitor Oracle Fusion Cloud Applications with Datadog Announcing the Datadog Terraform provider v4.0.0 Scaling Kubernetes workloads on custom metrics How to design cloud environments for AI-powered threat analysis
Bring high-performance observability to secure Kubernetes...
2025-07-30 · via Datadog | The Monitor blog

In Kubernetes environments, applications often communicate with the Datadog Agent to send telemetry data such as custom metrics via DogStatsD or traces through Datadog APM. How this communication takes place depends on the communication mode set on the Datadog Cluster Agent’s Admission Controller. With the sockets option, communication takes place through local inter-process communication via Unix domain sockets (UDS), whereas the service and default hostip options rely on network communication.

Among these methods, the sockets option via UDS is typically preferred because it bypasses the network stack entirely, resulting in better speed and performance. The sockets option also helps improve observability because UDS makes it easier for the Agent to identify the origin of packets sent to the socket. However, enabling UDS-based communication has historically required mounting socket files via hostPath volumes, a method incompatible with namespaces governed by non-privileged Pod Security Standards (PSS).

To resolve this issue, Datadog has released a CSI (Container Storage Interface) driver that mounts UDS sockets into pods by using CSI volumes instead of hostPath volumes. CSI volumes are compatible with all PSS levels, including the restricted level, making it possible for teams to adopt UDS-based observability features without violating security constraints.

In this post, we’ll look at:

Why hostPath volumes are blocked in secure environments

In Kubernetes, hostPath volumes allow pods to mount files or directories directly from the host node’s filesystem. Though flexible, this mechanism introduces significant security risks. For example, malicious or compromised pods could access sensitive host data—such as kubelet credentials—or interact with the container runtime in ways that could lead to privilege escalation or lateral movement.

For these reasons, Kubernetes defines hostPath as a privileged capability and disallows its use under the baseline and restricted Pod Security Standards. These standards are increasingly adopted in security-conscious organizations and are documented in the Kubernetes Pod Security Standards guide.

How UDS sockets are mounted securely with the Datadog CSI driver

The Datadog CSI driver provides a Kubernetes-native way to mount DogStatsD and APM UDS sockets into application pods using CSI volumes. Introduced as part of the CSI GA effort by Kubernetes (as explained in this blog post), CSI allows third-party vendors to implement secure volume plugins without relying on hostPath.

By using the CSI driver, Datadog shifts the hostPath volume from the user namespace to the CSI driver namespace. With this shift, user apps mount a Datadog CSI volume, and there is no need for user apps to mount a hostPath volume since it is replaced with a CSI volume. This new arrangement allows your services to use high-performance telemetry data and APM trace ingestion via sockets even in namespaces with restrictive pod security standards.

The CSI driver is deployed as a Kubernetes DaemonSet and is open source, with support for manual setup or Helm-based configuration. It is compatible with all CSI-enabled Kubernetes versions (v1.13+), and integrates directly with the Datadog Admission Controller to simplify adoption.

Diagram showing Datadog CSI driver mounting sockets via Kubernetes CSI volumes.

Which mount types are supported for observability sockets

The CSI driver supports four mount types, enabling both granular and directory-level socket access. You can configure:

  • DSDSocket: to mount only the DogStatsD socket file
  • APMSocket: to mount the trace agent’s socket file
  • DSDSocketDirectory: to mount the parent directory of the DogStatsD socket
  • APMSocketDirectory: to mount the parent directory of the trace agent socket

These options are defined in the type property of Datadog CSI volume’s VolumeAttributes and used in conjunction with standard pod volumeMounts. Example manifests and usage guides are available in the GitHub README.

How CSI volumes are injected using the Datadog Admission Controller

Datadog’s Admission Controller automatically configures observability communication for pods based on the selected mode: service, hostip, socket, or the newly added csi. When csi mode is enabled, the Admission Controller injects a CSI volume instead of a hostPath volume—ensuring compliance with any enforced PSS policies.

The admission controller handles this transformation transparently, as shown in the sequence diagram below. When a user creates a pod, the Admission Controller mutates the pod spec to include a CSI volume source for socket-based communication. When the pod is scheduled, the kubelet requests the CSI volume, and the Datadog CSI driver mounts the appropriate socket into the container.

Sequence diagram of Datadog CSI volume injection: The Admission Controller mutates the pod spec, and the CSI driver mounts the socket via the kubelet.

What’s next for CSI support

In upcoming updates, Datadog is planning to extend the CSI driver to support mounting APM Single Step Instrumentation (SSI) libraries via CSI. This will allow pods that use APM SSI to start faster and use disk more efficiently, as the libraries will be available at container startup without needing to be downloaded or injected post-launch.

Run UDS-based observability securely with Datadog CSI

With the Datadog CSI driver, Kubernetes users can enjoy the full performance benefits of UDS-based observability—even in environments governed by restrictive Pod Security Standards. The CSI-based approach ensures compliance with Pod Security Standards and integrates with Datadog Admission Controller to automatically inject the CSI volume into pods.

To learn more, check out the Datadog CSI driver on GitHub and read our documentation for step-by-step setup instructions. If you’re not yet using Datadog, you can sign up for a 14-day free trial.