Lab: Build Executive KPI Dashboard with Drill-Down

Create the dashboard. VCF Operations → Dashboards → Create Dashboard. Name='Infrastructure Health & Operations'. Layout: 3 rows. Row 1 (executive): 4 columns for scoreboards. Row 2 (correlation): 2 columns — heat map + trend chart. Row 3 (detail): 2 columns — object list + alert list.

Row 1 — Executive Scoreboards. Add 4 Scoreboard widgets: (a) 'Total VMs' — metric: count of powered-on VMs. Thresholds: none (informational). (b) 'Active Alerts' — metric: active alert count. Thresholds: green <5, yellow 5-20, red >20. (c) 'Avg CPU Usage' — metric: avg cpu.usage across all clusters. Thresholds: green <60%, yellow 60-80%, red >80%. (d) 'Capacity Days Remaining' — metric: min Time Remaining across all clusters. Thresholds: green >90 days, yellow 30-90, red <30. These 4 widgets give an executive everything they need in a single glance.

Row 2 — Heat Map and Trend Chart. Add Heat Map widget: Group By=Cluster, Size=VM Count, Color=CPU Usage (green=low, red=high). This shows which clusters are hot and how many VMs they contain — cluster 'prod-01' being red with 200 VMs is a different priority than 'dev-test' being red with 10 VMs. Add Metric Chart widget: configure as External-Provider (will be driven by heat map selection). Default metric: cpu.usage for all clusters over 7 days.

Configure widget interaction: Heat Map → Metric Chart. Edit the Heat Map widget → Output Filter → configure to send selected object to the Metric Chart widget. Now clicking a cluster in the heat map updates the trend chart to show that cluster's CPU/Memory/Storage usage over time. This is the drill-down pattern — executive sees red cluster → clicks → sees the trend → identifies when it started spiking.

Row 3 — Object List and Alert List. Add Object List widget: configure as External-Provider (driven by heat map selection). Columns: Object Name, CPU Usage, Memory Usage, Disk Latency, Alert Count. Sort by CPU Usage descending. Add Alert List widget: configure as External-Provider. Filter: Active alerts only. Sort by criticality descending. Both widgets update when a cluster is selected in the heat map.

Add secondary interaction: Object List → Metric Chart. Configure the Object List to also send selected objects to the Metric Chart. Now the user can: (1) click a cluster in heat map → see all hosts in that cluster in the object list → (2) click a specific host → see detailed metrics for that host in the trend chart. This creates a 3-level drill-down path: Cluster → Host → Metrics.

Add alert integration. Configure the Alert List widget: Interaction → clicking an alert opens the alert detail page (built-in behavior). Add a Text widget with static content: 'Dashboard Guide: Click cluster in heatmap → See hosts below → Click host → See metrics on right → Click alert for details'. This helps users who are new to the dashboard navigate the interaction model.

Test the full drill-down flow. Starting position: all widgets show summary data. Action 1: Click the hottest cluster in the heat map → Object List updates to show hosts in that cluster, Metric Chart shows cluster-level trends, Alert List filters to that cluster's alerts. Action 2: Click the busiest host → Metric Chart updates to host-level CPU/Memory/Disk trends. Action 3: Click an alert → navigate to alert detail with root cause. This end-to-end flow should take < 10 seconds.

Export dashboard for version control. Dashboards → select 'Infrastructure Health & Operations' → Actions → Export Dashboard. Download the JSON definition. Store in Git alongside super metric definitions from Lab vcap-ops-01. To deploy to another VCF Operations instance: Dashboards → Import → select JSON file. All widgets, interactions, and configurations are preserved.

Create a scheduled screenshot for NOC display. Configure a browser kiosk to load the dashboard URL in full-screen mode with auto-refresh (VCF Operations dashboards support URL-based auto-login with token). For NOC wall displays, use the 'Present' mode that hides navigation chrome and maximizes widget area. Set auto-refresh interval to 60 seconds for near-real-time updates.