Dashboard Overview Guide

Learn how to utilize the overall infrastructure status dashboard that appears after logging in to the GIIP platform.

🚀 Go to Dashboard →

📋 Overview

The Dashboard is an integrated control screen that allows you to grasp the status of your infrastructure distributed worldwide at a glance. It visualizes real-time metrics and cost data to facilitate swift decision-making.

🧭 Layout — Three Tiers (giip #2443/#2444/#2445/#2446)

The top of the dashboard answers only two questions first: "what do I have to decide right now" and "how is every project doing". Everything else is pushed below. The goal is that even with many projects, the summary still fits on one mobile screen.

① What you need to decide

Shows up to 5 issues that cannot move without a human decision.

• Scope: only issues whose status is NEEDS_DECISION or WARN.
• Stalled REVIEWs, gate reverts and 60-minute inactivity are re-judged automatically by the hourly bot, so they do not appear here (they belong to tier ③).
• When there is nothing to decide, you get a single "Nothing to decide right now" line instead of an empty card.
• Above 5 items, the top 5 are shown plus a "+N more" link to the issue list filtered by NEEDS_DECISION.
• Tapping a row opens that issue.
• The count in this tier is itself a quality metric — the better the automation works, the smaller it gets.

② Project status

One line per project you can access.

Column	Meaning
In progress	IN_PROGRESS issues
Waiting	READY + PENDING issues
Blocked	NEEDS_DECISION + WARN + IN_PROGRESS with no comment for 60+ minutes + gate escalation (3 strikes)

• Traffic light: 🔴 when blocked >= 1, 🟢 when blocked is 0 and in-progress + waiting >= 1, ⚪ when everything is 0 (shown as "—" instead of numbers).
• ⚠️ "Blocked" in tier ② is broader than tier ①. Tier ① is "what a human must decide"; tier ② blocked is project health.
• An issue matching several conditions is counted once (no double counting).
• An issue that is in progress but stalled for 60+ minutes appears in both In progress and Blocked — that is intentional, so "running but stuck" is visible on one line.
• Tapping a row selects that project; tapping it again clears the selection. This selection applies inside this dashboard only (tier ③ then shows that project only).
• The "Enter →" button at the right of the row is different. It changes the project of your login session, exactly like the project selector in the top bar, and reloads the page so the whole app (server list, issues, KVS, …) switches to that project.
  • On narrow screens the text is hidden and only the → arrow is shown. The behaviour is the same.
  • On the project you are already in, the button turns into a greyed-out ✓ and cannot be clicked — pressing it would only reload the page.
  • In short: tapping the row = narrow down this screen, "Enter" = actually move into that project.
• Closed (DONE) issues are excluded from every count.

③ Blocked issues

Every flavour of "this issue is stuck somewhere" is gathered into one section. It starts collapsed (always collapsed on mobile), and the count badge next to the title tells you how many there are without expanding.

• The same issue appears on exactly one row. Previously five separate widgets each listed it (measured: issue #2148 appeared in three widgets at once).
• Instead, every reason it is stuck is shown as a badge. Two reasons means two badges.

Badge	Meaning
Auto-completion suspected	Status is WARN — the completion verdict kept being reversed
Needs your decision	Status is NEEDS_DECISION
Gate cap reached	The automatic revert hit its 3-attempt cap and now needs a human
Gate reverted	Reverted at least once, but not yet at the cap
REVIEW stalled / PENDING stalled	No activity in that state for 7+ days (days shown in parentheses)
No PR	In REVIEW but no matching PR exists in any repository
Idle	IN_PROGRESS with no comment for 60+ minutes (minutes shown in parentheses)

• Badge colours have three levels — red means a human must act, amber means automation is retrying or the issue is stalling, grey is supporting information.
• Selecting a project in tier ② narrows this list to that project only. Tap the same row again to go back to all projects.
• Completed (DONE) issues never appear.
• The first 20 rows are shown, with "+N more" revealing the rest — nothing is discarded, only folded away.

③ Detail — the remaining sections

Below Blocked issues, every error, infrastructure and reference section is collapsible too. They all follow the same rules (count badge, grey + bottom when empty, remembered state).

Section	What the header badge counts	Initial state	Contents
Blocked issues	Issues that are stuck	Expanded on desktop only	See ③ above
SES Proposals	Action-needed (pending+failed+analyzing) count	uLevel≥70 only	Public proposals (csn=0) status summary + top 3 action items
Open issues	Total number of unfinished issues	Always collapsed	Issues in progress or waiting (25 / 50 / 100 per page)
Error logs	Unresolved errors	Collapsed	Total / resolved / unresolved / recent tiles + breakdown by source and severity
Suggested actions	Suggestions available	Collapsed	Review or delete servers that have not been checked in a long time
Server health statistics	Unhealthy servers	Collapsed	Project and server counts + healthy/unhealthy donut + list of abnormal servers
Domain expiry	Domains expiring within 30 days	Collapsed	Sorted by nearest expiry (already-expired count gets its own badge)
Projects & servers	Projects	Collapsed	Servers / healthy / unhealthy per project (most unhealthy first)
KVS activity summary	Distinct factors with activity	Collapsed	What was collected in the last 24 hours

• Why Open issues starts collapsed: it is the longest list on the dashboard, so leaving it expanded pushes every section beneath it off the screen. Even collapsed, the number beside the title is the total count, so you know how many there are without expanding.
• A header badge is not "rows currently on screen" — it is "the reason to open this section". Error logs shows unresolved errors rather than total errors; Server health shows unhealthy servers rather than total servers.
• If the lookup fails the badge shows … instead of 0, because writing 0 would falsely signal "nothing wrong".
• The Error logs section only appears for accounts with uLevel 70 or above.

Collapse / expand

• The count badge next to each section title stays visible when collapsed — collapsing never hides the number.
• Empty sections are never hidden; they are greyed out and moved to the bottom of the list, so sections with something to show always stay on top.
• On first visit: on mobile (viewport narrower than 768px) tier ③ is entirely collapsed, so tiers ① and ② fit on the first screen. On desktop only the first tier ③ section ("Blocked issues") is expanded.
• Collapse state is remembered per browser, and once you collapse or expand a section yourself, your choice wins over the defaults above.
• Where storage is blocked (private windows), the page still renders correctly; only the memory is skipped.
• Collapsed content is removed from the page and loaded on expand, which keeps the initial mobile render cheap.

🔍 Key Widget Descriptions

1. Real-time Asset Status

Displays the total number and status (Running, Stopped, Warning) of servers and databases currently under management in card format. Clicking a card navigates to the corresponding list page.

2. Global Infrastructure Map

Displays the physical locations where servers are situated on a map. If a failure occurs in a cluster within a specific region, the icon on the map will flash red.

3. Multi-cloud Cost Summary

Shows the estimated cumulative costs for the current month from connected cloud accounts like Azure and AWS by proportion (Pie Chart). You can check the increase/decrease trends compared to the previous month.

4. Recent Issues & Alarms

Lists the most recently occurred Critical and Warning issues. You can immediately navigate to the detailed root cause analysis page by clicking an issue.

🛠️ Tips

• Filtering: Use the [Select Project] filter at the top of the dashboard to focus monitoring on data from a specific project.
• Refresh: Dashboard data is automatically updated every minute. If a manual refresh is needed, click the [Refresh] icon in the upper right.

💡 Important Notes

• If a specific metric is displayed as Unknown, please check the [Agent Status] of the corresponding server first.

API Reference

For detailed API specifications of this feature, refer to the dedicated guide.

📘 API Reference Overview (v1.2)

Troubleshooting

Symptom	Cause	Resolution
A metric shows Unknown	The agent on that server is not responding	Check the server's [Agent Status] and restart the agent.
Dashboard data does not refresh	Auto-refresh (1-minute cycle) is delayed or the session expired	Click the [Refresh] icon in the upper right, or log in again.
Cost summary is empty	No cloud account (Azure, AWS) is connected	Verify the cloud account integration settings.
Only one project's data is shown	The [Select Project] filter at the top is applied	Change the filter to all projects to display all data.

---

Version: 1.2 Last Updated: 2026-09-15 Source: giipv3/public/help/dashboard.en.md