Add detection for Talos provider - #3410
Conversation
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 2c58385e4e
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
| | `gke-autopilot` | Cluster (DDA) | Annotation only | Full GKE Autopilot workload adaptation (volume, env var, path, image, and PriorityClass changes). See [Datadog Operator on GKE Autopilot][3] | `providers.gke.autopilot` | | ||
| | `windows` | Node (DAP) | Annotation only | Builds a Windows-compatible node Agent DaemonSet on the targeted Windows nodes: Linux-only containers, mounts, and security context are stripped, and a Windows base image and init config are applied. Available in Operator v1.30.0+ | None | | ||
| | `talos` | Cluster (DDA)[^talos-scope] | Annotation only | Drops host volumes that don't exist on Talos Linux nodes: `/usr/src` and `/lib/modules` (OOM Kill, TCP Queue Length), and `/etc/passwd`/`/etc/group` (Live Process Collection, Process Discovery, CWS, CSPM) | `providers.talos.enabled` | | ||
| | `talos` | Cluster (DDA)[^talos-scope] | Detection or annotation | Drops host volumes that don't exist on Talos Linux nodes: `/usr/src` and `/lib/modules` (OOM Kill, TCP Queue Length), and `/etc/passwd`/`/etc/group` (Live Process Collection, Process Discovery, CWS, CSPM) | `providers.talos.enabled` | |
There was a problem hiding this comment.
This row now advertises Talos detection, but the earlier “Automatic detection” section still says detection uses only node labels and its provider table omits Talos and status.nodeInfo.osImage. Consequently, users following the detection documentation can conclude that a Talos node resolves to default, despite the new behavior automatically applying Talos-specific volume removals. Update that section to document the OS-image signal and Talos result.
Useful? React with 👍 / 👎.
This comment has been minimized.
This comment has been minimized.
tbavelier
left a comment
tbavelier
left a comment
There was a problem hiding this comment.
🔥
nit: providers.md can slightly be updated to mention we look at node info (labels, system info, etc.), not only labels to determine the cluster provider ? what codex mentioned already below
![datadog-official[bot]](https://avatars.githubusercontent.com/in/1358509?s=60&v=4){kind=small}
There was a problem hiding this comment.
More details
The Talos node image check keeps label-based provider detection as its fallback. The capability checks remove only the host mounts that Talos does not provide.
🤖 Datadog Autotest · Commit 5d010be · What is Autotest? · @DataDog review to ask questions · Any feedback? Reach out in #autotest
5d010be to
f308d42
Compare
2 participants
What does this PR do?
CONTP-2028
Add automatic detection of Talos provider added in #3406. Detection is based on node
status.nodeInfo.osImagevalue and assumes Talos Kubernetes clusters consist of uniformly Talos nodes.Motivation
What inspired you to submit this pull request?
Additional Notes
Anything else we should know when reviewing?
Minimum Agent Versions
Are there minimum versions of the Datadog Agent and/or Cluster Agent required?
Describe your test plan
Same as #3406 but apply DDA without provider annotation.
Checklist
bug,enhancement,refactoring,documentation,tooling, and/ordependenciesqa/skip-qalabel