Dashboards

Dashboards are part of the free visibility tier. They run on an Apache Superset instance dedicated to your tenant, embedded in the Infralign app — nothing to install, no separate BI login. They refresh nightly from your own Azure billing and resource data, and populate after the first run, within 24 hours of connecting.

The screenshots below are from the NovaPulse demo estate — an anonymised sample tenant, not real customer data.

Click any image to enlarge.

The four persona boards

Each cost lane carries four boards, one per audience, so a CFO and a platform engineer read the same numbers through the view that suits them.

Executive — Cost & Savings. Latest complete month spend, month-so-far like-for-like, projected month-end and projected-versus-budget, spend trend by service category, top movers, realised commitment savings, Savings Plan and Reservation coverage %, and cost by country.

FinOps — Commitments & Savings. Reservation and Savings Plan coverage against target, realised commitment savings, and the commitment analysis a FinOps owner uses to decide what to buy next.

Engineering — Efficiency & Resources. Idle VMs (30-day rolling p95 CPU under 5%) and their spend, recoverable € from rightsizing shutdown and resize candidates with confidence levels, Azure Hybrid Benefit licensing, cost-versus-CPU scatter, daily spend per resource, what changed over the last 7 days versus the prior 7, cost anomalies (peak versus 7-day average), a resource drill-down, and a cost-by-resource-group treemap.

Data Ops — Quality & Freshness. Pipeline freshness and data-quality checks, so you can see when each source last landed and trust the numbers on the other three boards.

The app carries its own pages alongside the Superset boards — Dashboard, Waste, Recommendations, Resources, FinOps Audits, Chat, and BI — linking the same data into the audit and approval workflow.

Reading the freshness banner

Every board carries an as-of line. Cost, utilisation, and inventory each carry their own date, because they land from different sources at different times. Read the date next to a figure, not just the figure.

The daily Cost API lane runs about one day behind, so yesterday’s spend is the freshest complete number, and more than a day or two behind is worth a look. A STALE banner means a source has fallen further behind than expected; check it on the Data Ops board. On the FOCUS (monthly) lane, up to a month behind is normal rather than stale.

Cost terms

The boards use these terms precisely. Quoting the wrong one is the commonest reconciliation error.

Term	What it means	When to quote it
Billed (actual) cost	What the invoice charges: the amount Azure bills for usage in the period.	Reconciling against the Azure invoice.
Effective (amortised) cost	Reservation and Savings Plan purchase prices spread across the resources that consume them, rather than landing as a lump on the purchase date. The boards default to this.	Steering decisions.
Like-for-like month comparison	Two months compared over an equal day count, with provisional (not-yet-final) days excluded.	Any month-on-month comparison where the current month is incomplete.
Coverage %	The share of commitment-eligible usage that sits under a commitment. Savings Plan coverage is measured on a € basis, Reservation coverage on an hours basis, each per commitment pool.	Per commitment pool only. Do not average coverage across pools — the denominators differ.
Recoverable (avoidable) spend	Idle-VM cost, plus half of each resize delta, plus the Azure Hybrid Benefit licensing estimate. Low-CPU burstable B-series VMs are excluded by design, because low average CPU is their intended operating mode.	Sizing the opportunity before the audit quantifies it.

Savings tiles

Two savings numbers appear on the boards, measured against different baselines.

Tile	What it measures	Baseline
Realised commitment savings	What your reservations and savings plans have already saved, banked.	Pay-as-you-go
Indicative savings	The saving a recommendation would deliver if approved.	Retail (pay-as-you-go) price

Never add the two together — the baselines differ, so summing them double-counts.

Some tiles ignore the date filter by design: projected-versus-budget, realised and indicative savings, coverage, and cost-by-country each describe a fixed period or a current position rather than the filtered window.

The cost lanes

The Cost API lane feeds every board by default, built from the Azure Cost Details API at daily granularity, merging Actual and Amortized metrics. It is the fresher lane and the one almost every estate runs on.

Very large estates can add an optional FOCUS lane, built from native FOCUS 1.2 exports on a monthly cadence. The same four boards exist for it. When both lanes run, use the daily lane for the freshest movements and the FOCUS lane for the standardised, portable dataset — setup is in FOCUS exports and storage setup.

Refresh and filters

The pipeline runs once a night and syncs the boards before the working day, so they are current each morning without any manual export or refresh. Stage times and the 13-month history backfill are in data sources.

Every board supports the same filters:

• Subscription — one subscription or the whole estate; production estates fan out across 10+.
• Service — Azure service or service category.
• Region — a country or region.
• Charge period — the month or date range in view.

Exporting numbers

Chart-level CSV download is available from the BI view for any tile. Monthly report figures can be provided as a spreadsheet on request during the engagement. A self-serve export API is on the roadmap.

Read-only

Every board is built from read-only access to your billing and resource metadata. The platform never modifies tags, resources, or exports to improve its own reporting.

---

Next: Reports — the written monthly companion to these boards.