Recipe R05
The Console
An operations dashboard: live numbers, a month of errors with their causes, today's latency, activity, deployments and ⌘K.
Teams building an internal dashboard, admin panel or the overview page of a developer product.
ScenarioThe overview page of a deployment platform: key numbers, thirty days of error rate with the deploys and incidents that moved it, today's latency for one API, the day's activity, recent deployments, and a command palette.
The page, section by section
- 01
Navigation
Workspace SidebarOne indicator slides between rows, stepping inward for nested items, so where you are is always drawn.
- 02
Key numbers
Metric MorphOnly the digits that changed turn over, coloured by whether the change is good; error rate and latency are inverted.
- 03
Error rate, with causes
Annotated TimelineThirty days of error rate with deploys, incidents and flags pinned underneath; selecting one shades its effect.
- 04
Latency
Latency HeatmapEvery request of the day by time and duration, with p50, p95, p99 and the target drawn across.
- 05
Activity
Activity StreamTime has weight: gaps grow with the time between events, and repeated actions fold into one line.
- 06
Deployments
Focus TableSortable, keyboard-navigable, with the rows around the focused one receding; compact records on narrow screens.
- 07
Commands
Shortcut Palette⌘K opens a palette that runs the command and teaches its shortcut for next time.
Change first
The first things to replace to make this page yours.
- Wire the key numbers to your real metrics, with units and time ranges.
- Feed the timeline your real deploys, incidents and flag changes, written the way your team would explain them.
- Point the latency heatmap at one real endpoint and set its target to your SLO.
- Fill the command palette with the actions your users actually repeat, with their real shortcuts.
The build brief
Paste it into Claude, Cursor or ChatGPT with the component files, and ask for your own version: your name, your work, your product. It names each component and where to copy it from.
# Build brief: The Console An operations dashboard: live numbers, a month of errors with their causes, today's latency, activity, deployments and ⌘K. For: Teams building an internal dashboard, admin panel or the overview page of a developer product. Scenario: The overview page of a deployment platform: key numbers, thirty days of error rate with the deploys and incidents that moved it, today's latency for one API, the day's activity, recent deployments, and a command palette. ## Direction - Near-black (#09090b) with raised panels and hairline rings; colour is for status only: green ready, amber building, red failed, blue for deploys on the timeline. - Every chart answers a question: the timeline says why the error rate moved, the heatmap shows the whole distribution rather than one average. - Numbers change in place and say how they changed; nothing blinks, nothing pulses for attention. - ⌘K is the fastest way to do anything; the button in the header teaches the shortcut. - Geist for the interface, Geist Mono for numbers and keys. ## Page map 1. Navigation: Workspace Sidebar. One indicator slides between rows, stepping inward for nested items, so where you are is always drawn. 2. Key numbers: Metric Morph. Only the digits that changed turn over, coloured by whether the change is good; error rate and latency are inverted. 3. Error rate, with causes: Annotated Timeline. Thirty days of error rate with deploys, incidents and flags pinned underneath; selecting one shades its effect. 4. Latency: Latency Heatmap. Every request of the day by time and duration, with p50, p95, p99 and the target drawn across. 5. Activity: Activity Stream. Time has weight: gaps grow with the time between events, and repeated actions fold into one line. 6. Deployments: Focus Table. Sortable, keyboard-navigable, with the rows around the focused one receding; compact records on narrow screens. 7. Commands: Shortcut Palette. ⌘K opens a palette that runs the command and teaches its shortcut for next time. ## Change first - Wire the key numbers to your real metrics, with units and time ranges. - Feed the timeline your real deploys, incidents and flag changes, written the way your team would explain them. - Point the latency heatmap at one real endpoint and set its target to your SLO. - Fill the command palette with the actions your users actually repeat, with their real shortcuts. ## Components Copy each component's files from its page (Code tab) into your project before you start, and use them as given: change props and content, not internals. - Workspace Sidebar: https://tryvitrine.dev/components/workspace-sidebar (A SaaS app sidebar where one active indicator travels through the tree — down the list, into a sub-section and back out — so you can see where you moved from, not just where you are.) - Metric Morph: https://tryvitrine.dev/components/metric-morph (A live metric where only the place values that changed move.) - Annotated Timeline: https://tryvitrine.dev/components/annotated-timeline (A metric over time with the things that happened to it pinned underneath — deploys, incidents and feature flags, each on a stem up to the line.) - Latency Heatmap: https://tryvitrine.dev/components/latency-heatmap (Every request in a day, by when it happened and how long it took: time across, latency up a log scale, each cell shaded by how many requests landed there, with the p50, p95 and p99 traced through it and the target drawn across.) - Activity Stream: https://tryvitrine.dev/components/activity-stream (An activity log where time has weight: the space between two events grows with the time between them, so a busy half hour reads dense and a quiet afternoon reads as a pause.) - Focus Table: https://tryvitrine.dev/components/focus-table (A data table with depth of field: the row you're on stays sharp while the rows around it fade by distance, and the grid itself stays put.) - Shortcut Palette: https://tryvitrine.dev/components/shortcut-palette (A command palette that teaches: type a few letters, run the command, and if it has a keyboard shortcut a quiet line tells you once the action has happened.) ## Build - Next.js (App Router), TypeScript, Tailwind CSS. No extra UI libraries. - Write real content for this site: real names, dates, numbers and sentences. No lorem ipsum, no "Feature one". - Compose the sections in the order of the page map. Keep each component's own spacing; add page rhythm with consistent section padding. - Mobile first: every section must work at 375px wide without horizontal scrolling. - Respect prefers-reduced-motion everywhere; every Vitrine component already does. ## Before you ship - Run `/responsive-art-direction` (https://tryvitrine.dev/workshop/skills/responsive-art-direction): A dashboard has to become a single-column workspace on phones, with the sidebar as a menu. - Run `/design-system-consistency` (https://tryvitrine.dev/workshop/skills/design-system-consistency): Dashboards collect one-off cards and badges; this keeps them to one system. - Run `/interface-copy` (https://tryvitrine.dev/workshop/skills/interface-copy): Incident notes, empty states and command names are the dashboard's real content. - Run `/responsive-audit` (https://tryvitrine.dev/workshop/skills/responsive-audit): Charts and tables are the usual sources of horizontal overflow. - Run `/accessibility-pass` (https://tryvitrine.dev/workshop/skills/accessibility-pass): Charts need text alternatives, tables keyboard access, and the palette focus handling.
Direction
- Near-black (#09090b) with raised panels and hairline rings; colour is for status only: green ready, amber building, red failed, blue for deploys on the timeline.
- Every chart answers a question: the timeline says why the error rate moved, the heatmap shows the whole distribution rather than one average.
- Numbers change in place and say how they changed; nothing blinks, nothing pulses for attention.
- ⌘K is the fastest way to do anything; the button in the header teaches the shortcut.
- Geist for the interface, Geist Mono for numbers and keys.