Assess configuration performance and cost
Choose an observation window and cohort size, compare observed operation runtimes, and inspect the assumptions behind an optional cost estimate.
Before you start
- A verified operator, administrator, or owner session and the required workforce analytics permission.
- Operation timing enabled in agent policy.
- Role and configuration mappings supplied for the measured devices, with at least five people per configuration cohort.
Screenshots show the Rapport interface with example data. Use your organization’s devices and values. Select an image to view it at full size.
Open the assessment and review its inputs
Open Insights → Experience → Configuration cost, or search for Configuration cost. Read What it needs and expand How this is measured to check that your organization has the required observations and mappings.
Set a window, cohort floor, and optional hourly rate
Complete Assessment inputs. A useful first run is a 30-day observation window with the minimum cohort your organization permits. Leave cost blank if you only want the runtime comparison.
| Field or control | What to do |
|---|---|
| Observation window | Choose Last 14 days, Last 30 days, Last 60 days, or Last 90 days. |
| Minimum cohort | Choose 5, 10, 20, or 50 people. A cohort must meet the privacy floor to be returned. |
| Loaded hourly cost (USD, optional) | Enter the assumed USD hourly rate used to value excess operation runtime, for example 85. The accepted value is non-negative, no greater than 1,000,000, with at most four decimal places. |
Run the assessment
Click Run audited assessment and wait for the result. This reads aggregate observations for the chosen inputs and records the assessment. It does not deploy a configuration or change the endpoints.
What you should see: The page shows a retrieval time and the returned configuration comparisons, or a clear explanation if no eligible result is available.
Compare configurations with their reference cohort
For each returned role and configuration, inspect the cohort size, measurement count, observed operation duration, reference cohort, and excess runtime. Use the displayed baseline, image, ring, or fingerprint information to identify the configuration you are assessing.
- The reference is the fastest eligible cohort for that role in the assessed window.
- Excess runtime per person per week divides the observed excess by people and weeks in the observation window.
- An entered hourly rate produces an estimated value of the measured runtime difference. Use the displayed assumptions when sharing it.
Inspect mechanism evidence and coverage
Review Configuration-linked mechanism evidence when available. It groups conventional resource, stability, lifecycle, and control-plane network observations against the mapped role and usable configuration fingerprint. Read any missing-mapping, ambiguous-fingerprint, truncated-result, or suppressed-difference message before comparing cohorts.
Turn an observation into a small comparison
Identify a configuration difference worth investigating, inspect the contributing device and deployment information through the relevant operational guides, and plan a measured pilot. After collecting the follow-up observations, run a new assessment with the intended window and rate.
What you should see: You have a documented runtime comparison and its assumptions to inform the next investigation or pilot.
If something doesn’t work
No cohorts are returned
Check operation timing, identity and configuration mappings, and the selected cohort floor. Try a longer observation window when recent measurements are insufficient. Changing the floor cannot make missing mappings usable.
The cost or annual projection is absent
Enter an hourly cost to request a monetary estimate. For annual projections, inspect the mapping and scan-coverage requirements; missing or incomplete coverage prevents a projection.
The assessment reports an incompatible result or becomes stale
Run it again after checking the source and deployment state. Reconnect first if offline. After five minutes, the page asks for a new assessment before relying on those observations.