Operations Hub
- See incident analytics, automation health, and daily metrics in one dashboard — no switching between pages
- Three tabs: Overview (metrics and system health), Incident Analytics (hours saved, resolution rates, time-to-mitigate), Automation (task and trigger KPIs)
- Click any KPI card to drill into detailed charts and breakdowns
- Available for all agents — appears automatically in the sidebar
The problem: monitoring your agent means clicking through multiple pages
Your agent handles incidents, runs scheduled tasks, processes HTTP triggers, and consumes Azure Agent Units. But understanding whether your agent is performing well requires visiting multiple pages — incident metrics under one section, individual task pages for automation results, settings for consumption data, and chat threads for pending approvals.
There is no single view that answers "How is my agent doing?" You end up clicking through four or more sections to piece together the picture yourself.
How Operations Hub works
Operations Hub consolidates monitoring into three tabs accessible from the sidebar. Click Operations Hub in the left sidebar to open the dashboard.
Overview tab
The default view shows your agent's operational status at a glance:
| Section | What it shows |
|---|---|
| Status bar | Data source connection health — which connectors are configured and which need attention. Click Complete setup to configure missing sources directly. |
| Metrics chart | Daily Volume by Source, Daily Active Flow Consumption (AAU), or Daily Custom Tool Calls — select from the dropdown |
| Pending Actions | Count of conversation threads waiting for your input — approvals, questions, incident inputs, and pending command executions |
| System Health | Live agent process status and connector health |
Use the Time range selector at the top to adjust the reporting period (default: last 7 days). Click Refresh to load the latest data.
Incident Analytics tab
An interactive analytics dashboard with four KPI cards:
| Card | What it shows |
|---|---|
| Estimated engineering time saved | Total hours your agent saved, with a daily breakdown drill-down |
| Agent-supported resolution rate | Percentage of incidents resolved, with outcome breakdown: agent mitigated, human mitigated (agent assisted), human mitigated (no agent), and in-progress |
| Median Time to Mitigate | P50 comparison — agent resolution time vs. human resolution time |
| IntentMet evaluation score | Quality rating (1–5 scale with star visualization) based on how effectively the agent resolved incidents |
Click any KPI card to expand a drill-down view with detailed charts. Below the KPI cards, the Incident Volume & Outcomes chart shows incident trends over time.
Automation tab
A dashboard for monitoring scheduled tasks and HTTP triggers:
| Card | What it shows |
|---|---|
| Total Automations | Count of active automations, split by scheduled tasks and triggers |
| Total Runs | Execution count with trend comparison to the previous period |
| Success Rate | Percentage of successful runs with succeeded/failed counts |
| Avg Run Duration | Mean execution time with median and P95 percentiles |
Below the KPI cards, the Automation Overview section lists each automation with its type (Scheduled Task or HTTP Trigger), schedule, autonomy mode, run count, and success rate. Expand any card to see an AI-generated summary of recent executions.
Use the Search box and Type filter to find specific automations.
Before and after
| Before | After | |
|---|---|---|
| Checking agent health | Visit multiple pages — Monitor, Settings, individual tasks | Open Operations Hub — one dashboard, three tabs |
| Incident analytics | Navigate to separate metrics views, manually cross-reference | Incident Analytics tab with KPI cards and drill-down charts |
| Automation monitoring | Click into each scheduled task individually | Automation tab shows all automations with health indicators |
| AAU consumption | Navigate to Settings → Agent consumption | Overview tab metrics chart includes AAU data |
| Pending approvals | Scan chat threads for items waiting for input | Pending Actions count visible on the Overview tab |
Related capabilities
| Capability | What it adds |
|---|---|
| Track Incident Value → | Deep-dive incident analytics with per-response-plan performance, root-cause treemap, and drill-down to individual incidents |
| Monitor Agent Usage → | Detailed AAU consumption tracking and allocation management |
| Scheduled tasks → | Create automated workflows that appear on the Automation tab |