docs: add node-level scheduling policies for standalone workloads - #12
Merged
Conversation
There was a problem hiding this comment.
🟡 Changes recommended
There are a few objective documentation issues (English-only headings in the zh page and a grammar issue in the English page) that should be corrected before merging.
Once you've addressed the issues Copilot identified, you can request another Copilot review.
Pull request overview
Adds a new documentation page describing node-level scheduling policies for non-Kubernetes workloads (e.g., standalone/Docker vLLM on DGX Spark / GB10), and wires it into the existing MkDocs navigation with cross-links from the Kubernetes-focused GUI guide.
Changes:
- Add new Node Policies guides in English and Chinese (
node-policies.md,node-policies.zh.md) covering node selection, TID-aware/proc/.../commmatching, priority/time-slice semantics, API usage, and troubleshooting. - Add the new page to MkDocs navigation (including zh nav translation label).
- Add tips in existing GUI docs to redirect readers to the node-level policy guide when appropriate.
File summaries
| File | Description |
|---|---|
| mkdocs.yml | Adds the new Node Policies doc to nav and provides a zh nav label translation. |
| docs/node-policies.md | New English guide for node-level scheduling policies, including examples and troubleshooting. |
| docs/node-policies.zh.md | New Chinese guide for node-level scheduling policies, including examples and troubleshooting. |
| docs/gui.md | Adds a tip linking readers to the new node-level policy guide for non-Kubernetes workloads. |
| docs/gui.zh.md | Adds a similar tip in Chinese linking to the new node-level policy guide. |
Review details
- Files reviewed: 5/5 changed files
- Comments generated: 3
- Review effort level: Lite
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
| Current Gthulhu semantics are: | ||
|
|
||
| - `Priority > 0`: the strategy is **boosting**. The matching task receives priority treatment in the user-space scheduler. | ||
| - `Priority == 0`: the strategy is **non-boosting**. In user-space mode, it can still apply a custom time slice without jumping the run queue. |
| @@ -0,0 +1,211 @@ | |||
| # Node-Level Scheduling Policies | |||
|
|
||
| 對 vLLM policy 而言,也應確認 regex 真的匹配到預期的 TID。Policy 即使成功送到 node,如果 task name 沒有匹配,仍可能看不到預期的排程效果。 | ||
|
|
||
| ## Troubleshooting |
ianchen0119
approved these changes
Sep 7, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Add first-class documentation for node-level scheduling policies, with a focus on non-Kubernetes workloads such as standalone/Docker vLLM on DGX Spark / GB10.
What changed
node-policies.mdandnode-policies.zh.md/proc/<tgid>/task/<tid>/commmatchingPriority > 0boosting vsPriority == 0slice-only semanticsThe guide links to the live Node Policies mock at
https://gthulhu.github.io/Gthulhu/#/node-policiesinstead of embedding a screenshot, so the docs do not go stale as the mock UI evolves.Why
The existing docs describe Kubernetes-oriented scheduling policies, while node-level policies are the key path for host processes and non-Kubernetes inference servers. vLLM users running directly on DGX Spark should not need to move the workload into Kubernetes just to use Gthulhu scheduling.
Validation
Content is grounded in the current Manager NodePolicy API, Web UI fields, Decision Maker TID-aware matching behavior, and current plugin priority/time-slice semantics.