Utilization Data Collection
Overview
Inventory tells you what is in an environment. Utilization data tells you how hard it is working — CPU, memory, and (for some sources) disk activity sampled over time. Utilization is what turns a list of devices into a sizing recommendation: a server idling at 8% CPU and 20% RAM is a very different migration target than one running at 90% CPU around the clock.
Block 64 collects utilization data in two ways:
- Cloud integrations — for Azure and AWS environments, Insights polls the provider’s own monitoring service (Azure Monitor and Amazon CloudWatch) for each connected subscription or account.
- On-prem collectors — your Discovery Appliance, Discovery Application, and Discovery Agents collect utilization from the devices they scan. See Utilization Collection with On-Prem Collectors.
- The device is inventoried by a cloud integration but Collect Utilization Metrics is turned off. See Collect Utilization Metrics from Azure or Collect Utilization Metrics from AWS.
- The device is inventoried by an on-prem collector that has not yet delivered utilization data — collection runs on a schedule, and newly discovered devices take time to build a picture.
- The device was discovered recently and has not been sampled long enough for percentile figures to be meaningful.
- Azure — utilization metrics are read through Azure Monitor. Depending on your tenant’s scale and existing Azure Monitor usage, enabling collection can affect your Azure bill. See Azure Monitor pricing.
- AWS — utilization metrics are read through Amazon CloudWatch. See CloudWatch pricing.
- On-prem collectors — collection is performed by appliances and agents you already run; there is no additional provider cost.
- Collect Utilization Metrics from Azure
- Collect Utilization Metrics from AWS
- Utilization Collection with On-Prem Collectors
- Time Series Data Roll-ups
Where utilization data appears
Utilization data powers the Cloud Sizing reports in Insights (Reports > Cloud Sizing), including the Compute report. Each row shows utilization figures alongside the device’s inventory data:
|
Column |
What it shows |
|
95th CPU % |
CPU utilization at the 95th percentile — the ceiling of “normal” operation, ignoring short spikes |
|
95th RAM % |
Memory utilization at the 95th percentile |
|
Avg / Max CPU %, Avg / Max RAM % |
Optional columns — enable them from the table’s column chooser |
We lead with the 95th percentile rather than the average because averages hide peaks. A virtual machine that sits idle all week and runs flat out every Friday needs to be sized for Friday. [CONFIRM: state the aggregation window, e.g. “measured over the most recent 30 days of collected data.”]
Why some rows show “-”
The Compute report shows a dash (-) — not zero — when a device has no utilization data. Zero would mean “measured, and genuinely idle”; a dash means “no measurements yet.” Common reasons:
Insights flags these situations for you: if the report can tell that collection is available but switched off, a banner at the top of the report links straight to the relevant setting.
The Collector column
The Collector column on the Cloud Sizing Compute report shows how each device’s data was gathered — for example Appliance, Application, Agent (remote devices), or Integration for resources inventoried through your Azure or AWS connections.
Cost considerations
Privacy and footprint
Utilization collection gathers performance counters only — CPU, memory, and disk activity percentages. It does not inspect files, emails, or application content, and it does not install anything on the devices being measured beyond what your chosen collection method already uses.