Loading
Preparing this Costframe view
Preparing this Costframe viewTHE FINDINGS ENGINE
How a finding earns
its place on the list.
Severity, confidence, evidence — and totals you can trust. We don’t inflate savings; we provide engineering-grade truth.
Anatomy of a Detected Signal (Resource Performance Metric)
The Finding Anatomy
Structured, actionable data
A specific resource (or scope) with an estimated monthly saving at your net price, a recommended action, and the evidence behind it.
Severity Scoring
Impact vs Ease & Risk
Weighs financial impact against ease and risk of the change. High = large saving, low risk. Lower severity items are prioritized lower to reduce backlog noise.
Confidence Matrix
Evidence-backed analysis
Reflects how much evidence backs the call — e.g., sustained low CPU over a multi-week window. If data is sparse, confidence drops, preventing false alarms.
Audit Verification
Evidence & formula level detail
Each finding carries its data source, the window, the formula, and assumptions — so engineers can verify, not just trust. Full technical transparency.
Operational Review
Honest €0 Hygiene Candidates
Some items (orphaned configs, hygiene issues) carry no direct saving. We flag them separately as €0 candidates and keep them out of savings totals so nothing is inflated.
Trial Experience
Onboarding Preview
On-demand pilot scans let you preview the top three findings in detail; full resource-level detail and export capabilities unlock on the paid tier.
Evidence Ledger
A finding shows the proof and the doubt.
finding · vm_resizereviewable
Evidence checked
CPU history, cost line item, resource state, SKU catalog reference
Evidence unavailable
Missing metric windows, unsupported provider fields, stale billing exports
Decision state
Confidence, contradiction checks, owner, approval status, verification plan
- 01Conceptual diagrams are labeled as diagrams, not dashboards.
- 02Illustrative examples can explain math or workflow, but they cannot replace product proof.
- 03Every savings figure shown publicly stays illustrative until it is verified against your own audit.
Feature proof pack
Workflow, proof, limits, setup, and screenshot truth.
Workflow
Findings starts from provider evidence, moves through owner review, and remains customer-approved before action.
Proof
Public proof must cite checked evidence, unavailable evidence, freshness, confidence, and export or delivery state where relevant.
Limitations
Missing provider signals, readiness-gated providers, stale billing, and incomplete ownership lower confidence instead of being hidden.
Setup
Setup begins with read-only provider scope, validated connection health, and a first audit that exposes evidence coverage.
Screenshot state
No real screenshot is shown unless an approved anonymized product capture exists. Keep placeholders until then.
Architecture
Provider data flows into normalized evidence and review artifacts. Provider writes stay outside Costframe.
FAQ
Does this feature change cloud resources automatically? No. Costframe keeps provider access read-only and produces evidence, reports, drafts, or delivery records for human review.
Analyze your cloud cost posture
Connect your subscription read-only, trigger on-demand sweeps, and verify every single finding instantly.