Dashboard & Metrics

Orchestra exposes operational and AI consumption metrics at two levels: an in-app dashboard for a live operational overview, and the .NET Aspire dashboard for deep infrastructure and OpenTelemetry telemetry.

Dashboard

The Dashboard is the first screen after entering a workspace. It provides a high-level snapshot of everything happening across agents, tickets, and jobs — without navigating away.

Dashboard
Operations
🤖
Active Agents
0 / 10
Busy / total
Jobs Today
15
Success Rate (7d)
89%
17 completed · 2 failed
🕓
In Progress
0
Tickets
Total Tickets
19
Resolved (7d)
6
unique tickets
Code Impact (7d) CLI only
Added
+247
Removed
-89
Files
12
AI Consumption (7d)
View AI Usage →
Total Tokens
4.9M
API Calls
335
System
Agents
10
Skills
8
Workflows
2
Jobs Per Day (7d)
MonTueWedThuFriSatSun
Recent Jobs
All jobs →
Fix null ref in auth middleware
🎻 Maestro · PROJ-142
● Running
Add dark mode to dashboard
🎵 Aria · #304
Done
Payment gateway timeout error
🔍 Orpheus · PROJ-139
Done
Implement 2FA support
🎻 Maestro · PROJ-138
Failed
The Dashboard — live operational overview for the current workspace

Operations

The Operations row surfaces the most time-critical numbers at a glance:

MetricDescription
Active AgentsCurrently busy agents out of total configured agents (e.g. 0 / 10)
Jobs TodayNumber of jobs started within the current calendar day
Success Rate (7d)Percentage of completed jobs over the last 7 days, with a completed / failed breakdown
Jobs In ProgressJobs currently in a Running or Waiting state right now

Tickets

MetricDescription
Total TicketsAll tickets currently synced across every connected integration (Jira, GitHub, GitLab)
Resolved by Agents (7d)Unique tickets that an agent successfully acted on in the last 7 days

Code Impact CLI agents only

Tracks the concrete code changes produced by CLI-type agents (Coding Agent, Agentic Search Agent):

MetricDescription
Lines AddedCumulative lines of code added by CLI agents over the selected period
Lines RemovedCumulative lines of code removed by CLI agents
Files ChangedTotal distinct files modified across all CLI agent job runs

AI Consumption (7d)

A compact summary of AI API usage for the workspace, with a link to the full AI Usage page:

#
Total Tokens

Combined input + output tokens consumed by all agents in the last 7 days.

📶
API Calls

Total number of model inference requests made across all agents and jobs.

System

A quick inventory of the workspace configuration — how many agents, skills, and workflows are set up:

MetricDescription
AgentsNumber of agents configured in this workspace
SkillsNumber of skill folders created
WorkflowsNumber of automation workflows defined

Charts

📊
Jobs Per Day (7d)

A bar chart showing daily job volume over the last 7 days, helping you spot activity spikes or quiet periods.

🕓
Recent Jobs

A live list of the most recently run jobs with status badges and direct links to their detail views.

AI Usage

The AI Usage page gives a detailed breakdown of model consumption across the workspace. Navigate here from the Dashboard via View AI Usage → or through the workspace navigation.

← Dashboard
AI Usage
24h 7d 30d 90d All
#
Total Tokens (7d)
4.6M
#
Input Tokens (7d)
4.5M
#
Output Tokens (7d)
64.2K
📶
API Calls (7d)
335
Lines Added (7d)
+247
Lines Removed (7d)
-89
FILTER: Provider 1 Model ▾ Clear filters
Tokens by Model
claude-haiku-4.5
sonnet
gpt-5-mini
Usage Timeline (tokens/hour)
14h15h16h17h18h19h
The AI Usage page — token consumption, API call volume, and per-model breakdown

Summary Cards

The top row provides workspace-wide totals for the selected time window (24h / 7d / 30d / 90d / All):

CardDescription
Total TokensCombined input + output token count for the period
Input TokensTokens sent to the model (prompts, context, tool schemas)
Output TokensTokens generated by the model (responses, tool call arguments)
API CallsTotal number of model inference requests made in the period
Lines AddedLines of code added by CLI agents during the period
Lines RemovedLines of code removed by CLI agents during the period

Filters

Narrow the data to a specific subset using the filter bar:

🏢
Provider filter

Filter metrics by AI provider — Azure OpenAI, Ollama, or GitHub Copilot CLI — to compare cost profiles side by side.

🧠
Model filter

Drill into a specific model deployment (e.g. claude-haiku-4.5, gpt-4o-mini) to see its individual consumption share.

Charts

📊
Tokens by Model

A bar chart grouping total token usage by model. Quickly identify which model is responsible for the bulk of consumption.

📈
Usage Timeline (tokens/hour)

A line chart showing token throughput over the selected period. Useful for identifying peak usage windows and correlating spikes with specific jobs.

💡

Time window selector: Use the 24h / 7d / 30d / 90d / All buttons in the top-right corner of the AI Usage page to change the reporting window. All summary cards and charts update in sync.

Aspire Dashboard & OpenTelemetry

In addition to the built-in metrics pages, Orchestra ships full OpenTelemetry instrumentation for every .NET service. When running under .NET Aspire, all telemetry is automatically routed to the Aspire dashboard.

🔭
AI metrics via OpenTelemetry

The .NET Aspire dashboard surfaces AI usage metrics — token counts, model latency, and request volume — collected via OpenTelemetry from the API and Worker services.

🔍
Distributed tracing

End-to-end traces span from the API through the Worker service and into external AI provider calls, making it straightforward to diagnose slow or failing agent executions.

📋
Structured logs

All services emit structured logs captured by the Aspire dashboard, searchable and filterable by service, severity, and trace ID.

📡
Service health & resources

Container health, port bindings, environment variables, and resource graphs for every service — API, Worker, PostgreSQL, Redis, and Node.js services — are visible in the Aspire console.

💡

Accessing the Aspire dashboard: When you run aspire run, the Aspire dashboard starts automatically. The URL is printed in the console output (typically http://localhost:15888). No additional configuration is needed — all services register with the Aspire telemetry collector automatically.