Technical Overview
Architecture
The VCF Servers Insight adapter is a Go-based containerized collector that runs on a VMware Aria Operations Cloud Proxy. It communicates with Dell iDRAC / HPE iLO (Redfish) and Dell OpenManage Enterprise (OME) over HTTPS and pushes collected data to Aria Operations through the collector framework.
┌─────────────────────┐ Redfish / OME HTTPS ┌──────────────────────┐
│ Aria Operations │ ◄────────────────────────── │ iDRAC / iLO / OME │
│ Cloud Proxy │ /redfish/v1/... | /api/... │ (server platforms) │
│ ┌───────────────┐ │ └──────────────────────┘
│ │ VCF Servers │ │ HTTPS :443
│ │ Insight │ │ ◄──────────── register.indevops.com (image pull)
│ │ Adapter │ │
│ │ (Go container)│ │
│ └───────────────┘ │
└─────────────────────┘
Authentication
| Interface | Method |
|---|---|
| Dell iDRAC | HTTP Basic authentication |
| Dell OME | HTTP Basic authentication |
| HPE iLO | Session-based (X-Auth-Token via /redfish/v1/SessionService/Sessions/) |
A single shared credential set (username / password) is used; per-host credentials are resolved for multiple idrac_host entries.
Management Paths
Dell iDRAC (Redfish)
Connects directly to each iDRAC, collects systems, chassis, managers, and per-system metrics:
- CPU/memory utilization from Telemetry MetricReports (
/redfish/v1/TelemetryService/MetricReports/...) — dynamic report discovery - CPU, memory, storage, network, BIOS, thermal, power from Redfish endpoints
- Relationships: Adapter → Physical Server; (external) vSphere HostSystem → Physical Server
Dell OpenManage Enterprise (OME)
Connects to the OME console, fetches the device list, and for each server also connects to the blade's iDRAC (via IP from deviceManagement):
- OME-aggregated: temperature, power, uptime, inventory (fans, PSU, OS, NIC)
- Per-blade iDRAC Redfish: CPU/memory/storage/network/BIOS
- Network adapters use the OME
serverNetworkInterfaces(port/partition model) - Network switches use OME
portInformation(per-port metrics) - Relationships: Adapter → Chassis/Server; Chassis → Server/Switch; (external) vSphere HostSystem → Physical Server
HPE iLO (Redfish)
Connects directly to each iLO with session auth:
- CPU/memory utilization from
Oem.Hpe.SystemUsage(CPUUtil,MemoryBusUtil) — Intel only (AMD/ARM do not expose telemetry) - Uptime from
PowerOnMinutes - Storage, network, BIOS, thermal, power from Redfish
- Relationships: Adapter → Physical Server; (external) vSphere HostSystem → Physical Server
Redfish API Endpoints (iDRAC / iLO)
| Endpoint | Method | Purpose |
|---|---|---|
/redfish/v1/Systems | GET | List systems |
/redfish/v1/Chassis | GET | List chassis |
/redfish/v1/Managers | GET | BMC/firmware info |
/redfish/v1/Chassis/{id}/Thermal | GET | Temperature + fans |
/redfish/v1/Chassis/{id}/Power | GET | Power + PSUs |
/redfish/v1/Systems/{id}/Processors | GET | CPU |
/redfish/v1/Systems/{id}/Memory | GET | DIMM |
/redfish/v1/Systems/{id}/Storage | GET | Storage controllers |
/redfish/v1/Systems/{id}/EthernetInterfaces | GET | NICs |
/redfish/v1/Chassis/{id}/NetworkAdapters | GET | NIC model/firmware (Dell) |
/redfish/v1/Systems/{id}/BaseNetworkAdapters | GET | NIC model/firmware (HP) |
/redfish/v1/TelemetryService/MetricReports | GET | Metric reports (CPU/memory) |
/redfish/v1/TelemetryService/MetricReports/{id} | GET | Report content |
/redfish/v1/SessionService/Sessions/ | POST/DELETE | iLO session login/logout |
OME API Endpoints
| Endpoint | Method | Purpose |
|---|---|---|
/api/DeviceService/Devices | GET | List devices |
/api/DeviceService/Devices({Id})/Temperature | GET | Temperature |
/api/DeviceService/Devices({Id})/Power | GET | Power |
/api/DeviceService/Devices({Id})/SystemUpTime | GET | Uptime |
/api/DeviceService/Devices({Id})/InventoryDetails | GET | Inventory (fans, PSU, OS, NIC, ports, FRU) |
Collection Cycle
Collection is incremental and server groups are collected in parallel (errgroup, limit 5) for iDRAC and OME to keep within the collection window.
- Select interface — iDRAC / OME / iLO
- Authenticate — basic (iDRAC/OME) or session (iLO)
- Fetch systems/devices — systems list (Redfish) or device list (OME)
- Collect per server (parallel) — CPU/memory/telemetry, temperature, power, storage, network, BIOS
- Collect chassis — temperature, power, fans, PSUs
- Collect network switches (OME) — port count, active ports, per-port detail
- Build objects — properties and metrics per resource kind
- Build relationships — containment and external (vSphere)
- Logout / cleanup (iLO)
Non-fatal errors are logged and skipped; a partial collection is always returned.
Dashboards
The Management Pack ships three dashboards:
| Dashboard | Widgets | Purpose |
|---|---|---|
| VCF Servers Insight | Object List, Adapter Performance Scoreboard, Metric Picker, Metric Chart | Adapter overview, collected objects, adapter CPU/memory, ad-hoc metric analysis |
| Physical Servers Performance | Servers Performance Scoreboard, Object Relationship, Servers List | Per-server performance metrics, health/BIOS/BMC/model columns |
| Physical Servers Health | Health Heatmap, Active Alert List | Health status overview and only active alerts |
Object Model
Physical Server
- Identifiers:
name,serial - Groups/properties: General (manufacturer, model, uuid, power state, health, BIOS, ESXi version, CPU/memory usage), CPU, Memory, Storage Controllers, Physical Disks, Network Adapter, Chassis, Power Supply, Fan, BIOS
- Metrics: CPU Utilization, Memory Utilization, System Uptime, Temperature, Power Consumption, link speed/status
Chassis
- Identifiers:
name,serial - Groups/properties: General (temperature, power, model, health), Power Supply, Fan
- Metrics: Temperature, Power Consumption, fan speed, PSU output/input
Network Switch (LAN)
- Identifiers:
name,service_tag - Groups/properties: General (Port Count, Active Port Count, model, manufacturer, serial, MAC, firmware, IP), Network Port (per active/Up port: status, speed, role, MTU)
- Metrics: Port Count, Active Port Count, Port Speed Mbps
Alert Definitions
The pack ships 34 alert definitions covering:
- CPU utilization (warning ≥50%, critical ≥80%)
- Memory utilization (warning ≥50%, critical ≥80%)
- Temperature (warning ≥75°C, critical ≥85°C)
- Fan, power supply, chassis, server, network switch, NIC, physical disk, DIMM, storage controller health states
- BMC connection lost / server/chassis power off
Relationships
- Containment: Adapter → Chassis/Server; Chassis → Server/Switch (OME)
- External: vSphere HostSystem → Physical Server (match by HostSystem name)