> For the complete documentation index, see [llms.txt](https://uclix.gitbook.io/run-ai-rca-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://uclix.gitbook.io/run-ai-rca-docs/readme/ui-direction.md).

# UI Direction

> **Lens:** How it's built (UI) — the design contract for the operator console. **In this doc:** visual system · page model · interaction notes.

Run:AI RCA should feel like KubeRCA evolved for NVIDIA Run:ai operators.

**Who this is for:** designers and engineers shaping the operator console. The screen should help someone answer three questions quickly: what happened, what evidence supports it, and what safe next check should they run.

```mermaid
flowchart TB
  S[Incident list] --> D[Incident detail]
  D --> R[Problem, cause, actions]
  D --> E[Evidence trail and gaps]
  D --> C[Comments, feedback, re-analyze]
```

## Visual System

* Base: white and near-white surfaces.
* Accent: NVIDIA green `#76B900`.
* Text: graphite and charcoal, with green used for active state and healthy signals.
* Warnings: amber and red only for severity and error states.
* Layout: dense, scan-friendly, dashboard-first.

## Page Model

No separate agent-analysis pages are allowed in the MVP. Incident and Alert detail pages must contain:

* Entity header: severity, status, cluster, project, queue, workload, node.
* RCA summary: concise final answer.
* RCA body: root cause, impact, evidence, action items, missing data, prevention.
* RCA report: a structured, scannable report view, exportable as a styled Word (`.docx`) incident report with Korean-capable fonts, GFM tables, and a page footer carrying the incident id and page numbers.
* Similar Incidents: past look-alike incidents from dense RCA-content matching, re-ranked by family and workload identity. The query carries this incident's own RCA once it has one, so the panel compares diagnosis to diagnosis rather than alert title to alert title; before analysis lands, alert identity is all there is. Pod names are excluded on both sides — they never repeat across occurrences, so they can only add noise.
* Agent Evidence Trail: collector tabs in a single panel on the same page.
* Raw artifact viewer: collapsed by default.
* Chat panel: context-aware and tied to the current route.

## Interaction Notes

* The first screen is the dashboard, not a landing page.
* The user should be able to move from dashboard to detail and back without losing filters.
* The incident list is ordered newest-activity-first, and the timestamp it is ordered by is on screen as its own sortable column. An order whose key the operator cannot see reads as arbitrary.
* Realtime analysis updates should be visible via status badges and SSE events.
* Missing integrations should be explicit but not noisy.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://uclix.gitbook.io/run-ai-rca-docs/readme/ui-direction.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.