This release makes Global View the operating picture for cost and performance across AI and hybrid infrastructure. You get a Token Usage Dashboard that ties usage to latency and errors, a GPU Fleet Analysis dashboard across on-prem and cloud, Service Observability alerts inside Global View, deployment on VMs where needed, and more granular metering. Global View also correlates with new capabilities in Container Observability (GKE Autopilot, FSx for NetApp ONTAP, Azure Confluent), Infrastructure Observability (Dell PowerProtect, Dell ObjectScale), and Service Observability (Cisco Meraki C2C, NetFlow, advanced Kubernetes drilldowns, dashboard, and reporting upgrades).
Token Usage Dashboard for AI Workloads
How This Helps You
Puts token consumption next to latency and errors so you can see what is driving cost and where reliability is slipping.
Key Capabilities
- Per-service and endpoint token usage with time alignment to performance metrics
- Error class and latency percentiles alongside usage
- Optional cost mapping for dollars per set of tokens
- CSV and PDF export for audits and reviews
Benefits
- Clear link from AI workload usage to cost and user impact
- Faster detection of anomalous bursts tied to code or traffic changes
- Better input for budget planning and usage policies
How to Get Started
- Enable token data collection for the services you want to track
GPU Fleet Analysis Dashboard
How This Helps You
- Shows power, temperature, utilization, memory, and spend in one place to prevent throttling, optimize GPU utilization, and reduce idle capacity.
Key Capabilities
- Per-GPU power and thermal telemetry across on-prem and cloud
- Utilization, memory pressure, and clock metrics with node and cluster context
- Cost attribution by cluster, tag, or node group
- Indicators that point to likely throttling or configuration issues
Benefits
- Early warning on devices trending toward thermal limits
- Evidence to consolidate or repurpose underused nodes
- Justification for scaling decisions based on utilization and cost
How to Get Started
- Connect GPU telemetry sources for each environment
- Tag clusters and nodes for cost allocation
- Create fleet views by environment or workload and set notifications for heat and power thresholds
Service Observability Alerts in Global View
How This Helps You
Lets you review Service Observability alerts alongside container, infrastructure, and third-party signals without switching tools.
Key Capabilities
- Ingest of Service Observability alerts with source, severity, and tags
- Filtering by service, cluster, namespace, or device
Benefits
- Shorter triage time when multiple systems report similar symptoms
- Fewer handoffs and less guesswork during incidents
- One audit trail for post-incident reviews
VM-Based Deployment Option
How This Helps You
Supports organizations that standardize on virtual machines and are not ready to deploy new control planes in containers.
Key Capabilities
- Full Global View install on a VM footprint
- Same data ingestion and dashboards as container-native deployments
- Consistent RBAC and export options
Benefits
- Lower onboarding friction in VM-centric environments
- Easier compliance with existing platform standards
- Clear path to container-native adoption later without changing workflows
How to Get Started
- Select the VM deployment package for your hypervisor
- Follow the install guide to connect existing data sources
- Connect data sources like IO, CO, SO, or third-party observability/monitoring tools.
- Validate dashboards against a known incident before wider rollout
Granular Meter Cards by Hostname
How This Helps You
Reflects real cost differences between hosts, so mixed environments are reported accurately.
Key Capabilities
- Meter card definitions by hostname, in addition to model and location
- Cost and usage widgets that use per-host economics
- Roll-ups that respect host-level variation
Benefits
- More accurate showback and capacity planning
- Cleaner cost views when some hosts are premium or specialized
- Less manual reconciliation during reporting
How to Get Started
- Create or update meter cards at the hostname level
What to read next
Launch overview blog: How these updates improve performance visibility and ROI across AI workloads and hybrid infrastructure
Container Observability technical blog: GKE Autopilot support, FSx for NetApp ONTAP monitoring, Azure Confluent monitoring, multi-tenant tracking, and export enhancements
Infrastructure Observability technical blog: Dell PowerProtect and Dell ObjectScale integrations with topology and metrics
Service Observability technical blog: Cisco Meraki C2C, NetFlow, advanced Kubernetes drilldowns, and dashboard/reporting upgrades
If you want to see Global View in action with your workloads or discuss deployment options, contact your Virtana sales representative or request a demo.
Meeta Lalwani
