Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/admbe/FluxOp/llms.txt

Use this file to discover all available pages before exploring further.

The Flux overview dashboard is the operational home screen for your Azure estate. It shows your resource footprint by type and region, month-to-date actual and amortized spend with explicit scope reconciliation, telemetry coverage across your VM fleet, and the risk-adjusted dollar total of every actionable opportunity Flux has found. All values derive from the same governed data the rest of Flux uses — nothing is estimated or interpolated for display purposes.

Estate summary

The estate panel renders resource counts by type and by region in chart series so you can see the composition of your environment at a glance. Subscription-level breakdowns let you attribute footprint to specific billing scopes without leaving the dashboard. Chart series are built from the current resources_current view — the newest complete ARG snapshot for each configured subscription. If a subscription has not yet completed its first sync, it does not appear in the totals.
When Flux is running in snapshot mode, the data currency chip in the dashboard header shows the snapshot version that is currently loaded. Each analytical read comes from that pinned snapshot rather than a live database write.

Cost summary

The cost panel shows month-to-date spend in two tracks:
TrackSourceScope
Actual costCost Management Query API or FOCUS exportOn-demand, reservation, and savings plan charges as billed
Amortized costCost Management Query API or FOCUS exportCommitment charges spread across the reservation or savings plan period
Flux performs explicit current / history / commitment scope reconciliation before reporting a total. A subscription appears in the cost summary only after its cost data has been successfully collected; subscriptions still warming up are listed separately with a coverage note rather than assumed to be zero. FOCUS export data takes precedence over Query API data for any period where a manifest has been successfully ingested. This means the actual and amortized totals on the dashboard reflect charge-level FOCUS evidence where available, and Query API data for the remainder.

Telemetry coverage

Every VM in the estate receives an explicit telemetry coverage state. The dashboard coverage chip aggregates those states across your fleet:
StateMeaning
no-dataNo telemetry observation has been collected for this VM
warming-upTelemetry is present but below the minimum evidence window or coverage threshold
partialCPU is covered but network or memory metrics are missing
conflictingAzure Monitor and LogicMonitor disagree materially on CPU utilization
actionableFull multi-metric evidence is available for classification
The coverage chip on the dashboard reflects the proportion of VMs in an actionable or better state. Low coverage means right-sizing and idle-detection results apply to only part of your fleet; expanding Azure Monitor or LogicMonitor collection improves this without any Flux configuration change.

Valued action total

Flux calculates a single risk-adjusted monthly opportunity value — the number shown in the dashboard opportunity chip — by aggregating the individual valuations across every finding in the current opportunity queue.

Gross value

The estimated monthly saving if every finding were acted on immediately, using retail pricing for VM resize targets or the Advisor-supplied estimate as a fallback.

Risk-adjusted value

Gross value weighted by each finding’s confidence score. Low-confidence findings contribute proportionally less to the total, so the headline number stays honest.
Only findings where Flux can quantify a saving are included. Governance and review-only findings (for example, missing allocation tags or AHB eligibility reviews) are counted in the opportunity queue but contribute zero to the valued total unless a resource-level cost can be attributed.

Stale data warnings

Flux shows a staleness banner when collected data is older than expected, but only after the scheduled collection window has passed and a grace period has elapsed. This is schedule-aware staleness: a sync that has not yet been attempted today is not stale. The banner appears when:
  1. The last successful sync for a source (inventory, Advisor, cost, or telemetry) completed before the expected next scheduled run time, and
  2. A configured grace window beyond that expected time has also passed without a new successful run.
This approach prevents false warnings during the brief window between a scheduled trigger firing and its completion. If a sync is currently in progress, the dashboard shows an in-progress indicator rather than a staleness warning.
Ad-hoc sync via Integrations → Synchronize now covers inventory, Advisor, and Flux Intelligence. It deliberately excludes Cost Management to avoid competing for the shared QPU quota that the daily cost jobs also use.

Build docs developers (and LLMs) love