Managed FinOps for Azure, run by DynamoNet

Azure cost, mapped
to your products.
Savings found,
checked, delivered.

We run FinOps for your Azure estate. AI proposes a product for each resource, keeps unassigned cost visible and finds savings with the evidence attached. An AI critic and a named engineer check every finding. You approve each change and get a monthly Excel report.

Get a free read-only scan Sample report (.xlsx)

Read-only access. Nothing changes without your approval.
For Azure estates spending 20k or more a month.

✦ MAP · INVESTIGATE · CHALLENGE · REPORT

Dynamo · Northwind CareAgents working

Overview

Read-only

Project map

Resources clustered by project, dots sized by cost

30 resources

Claims $26.6KCare $20.6KMember $8.4KData Platform $23.5KShared $12.6KUnassigned $1.2K

AI project mapping

Resource to project, component and environment

27 of 30 mapped, 3 unassigned

ca-claims-api-prodContainer Apps

Claims › Claims API prod

High

aks-care-prodAKS

Care › Scheduler prod

High

vmss-ingest-prodVM Scale Sets

Data Platform › Ingestion prod

High

func-census-workerFunctions

Claims › Census Worker prod

Low

asp-member-devApp Service

Member › Member Portal dev

Medium

Evidence: names, tags, pipelines, App Insights dependencies, identities and traffic. Low confidence goes to the owner to confirm.

Recommendations

Investigated by AI · challenged by an AI critic

Savings, code total$10,470 to $13,910 / month2 checks to do first · $12.4K / month at stake

Compute savings plan across vmss-ingest-prod and aks-care-prodData PlatformCareMedium confidenceMedium risk

$11.7K/mo at stakeCheck first

Reserve 1-year capacity for sql-claims-prod and sql-providers-prodClaimsCareHigh confidenceLow risk

$2.5K to $3.2K/moSaving

Stop aks-care-staging node pools nights and weekendsCareHigh confidenceLow risk

$1.7K to $2.1K/moQuick win

Downsize asp-claims-prod from P2v3 to P1v3ClaimsHigh confidenceHigh risk

No savingRejected by critic

⌑

Read-only accessScoped reads, revocable by your admin.

✦

AI investigates. An AI critic challenges.Two independent agents on every finding.

⌁

Code checks. An engineer signs.Costs come from billing data; every saving is checked against it.

WHAT YOU GET

Three things.
Every month.

Cost by product, not by resource group. Savings with the evidence and risk attached. A report your CFO and your engineers can both open.

AI PROJECT MAPPING

Your resources, mapped to
product, component, environment.

Resource groups and tags tell part of the story. The mapping agent reads names, tags, pipelines, App Insights dependencies, identities and traffic, then proposes where each resource belongs, with a confidence level and the evidence. In the report this grouping is called a project: the product or application a resource serves.

High, medium or low confidence, with evidence, on every mapping
Owners confirm or correct; unsupported mappings stay unassigned
Shared services split by requests or ingested GB

⊞

Project mapping

98.7% of cost mapped

Proposed by the mapping agent. Confirm what is right, correct what is not.

Claims 28.6%Care 22.2%Member 9.1%Data Platform 25.3%Shared 13.5%Unassigned 1.3%

Claims

4 of 8 resources shown

Proposed

asp-claims-prodClaims Portal · prod · Tag app=claims; deployed by the claims-portal pipeline; App Insights dependency to ca-claims-api-prod

High confidence

ca-claims-api-prodClaims API · prod · Observed calls from Claims Portal (App Insights); same release pipeline

High confidence

redis-claims-prodClaims API · prod · Only client is ca-claims-api-prod (observed connections)

Medium confidence

func-census-workerCensus Worker · prod · Name similarity; managed identity writes to st-claims-docs

Low confidence

Waiting for the product owner

Care

4 of 5 resources shown

Proposed

sql-providers-prodProvider Directory · prod · Observed dependency from scheduler pods

High confidence

aks-care-prodScheduler · prod · Helm release labels app=care; namespaces scheduler and notify

High confidence

redis-care-stagingScheduler · staging · Only client is aks-care-staging

Medium confidence

func-notify-prodNotifications · prod · Identity access to the Care key vault only; no observed traffic yet

Low confidence

Waiting for the product owner

RECOMMENDATIONS WITH ESTIMATED SAVINGS

Savings you can approve,
because you can see why.

The investigator agent looks at patterns, not averages: peaks, memory, schedules, commitments and licences. An independent AI critic challenges every finding and rejects what the evidence does not support. The AI proposes a saving range, code checks it against the recorded cost, and an engineer signs it off.

A monthly saving range per item, with checks listed apart
Risk, confidence, evidence and rollback on every item
Realized savings measured after each change

✦

Recommendations

October 2026

Savings, code total$10.5K to $13.9Kper month, overlaps counted once

Checks to do first2$12.4K per month at stake

Compute savings plan across vmss-ingest-prod and aks-care-prodData PlatformCareCommitmentCritic: needs evidenceMedium riskMedium confidence

$11.7K/mo at stakeCheck first

Reserve 1-year capacity for sql-claims-prod and sql-providers-prodClaimsCareCommitmentCritic: passedLow riskHigh confidence

$2.5K to $3.2K/moAwaiting your approval

Stop aks-care-staging node pools nights and weekendsCareScheduleCritic: passedLow riskHigh confidence

$1.7K to $2.1K/moAwaiting your approval

Move container stdout logs to the Basic logs planSharedLogsCritic: passedLow riskHigh confidence

$0.9K to $1.1K/moAwaiting your approval

Apply Azure Hybrid Benefit to the warehouse SQL licenceData PlatformLicensingCritic: passed with a prerequisiteLow riskMedium confidence

$2.0K to $3.3K/moNeeds your information

Downsize asp-claims-prod from P2v3 to P1v3ClaimsRightsizingCritic: rejectedHigh riskHigh confidenceAverage CPU is low, but peak working set is 11.4 GiB, above the 8 GiB of P1v3

No savingRejected by critic

THE MONTHLY REPORT

One Excel workbook.
Every number traceable.

Each month you receive a workbook and a walkthrough with your engineer. Cost by product and service, the trend, every recommendation with its status, the mapping and the evidence gaps that still need your team.

Summary
Recommendations
Cost by project
Cost by service
Monthly
Resources
Project mapping
Evidence gaps

Download the sample workbook See what is inside →

XLSXdynamonet-sample-report.xlsx23 KB

IDRecommendationSaving / monthStatus

R-01Reserve 1-year capacity for sql-claims-prod and sql-providers-prod$2.5K to $3.2KAwaiting your approval

R-02Delete three unattached managed disks$0.3K to $0.4KApproved

R-03Stop aks-care-staging node pools nights and weekends$1.7K to $2.1KAwaiting your approval

R-04Move container stdout logs to the Basic logs plan$0.9K to $1.1KAwaiting your approval

R-05Apply Azure Hybrid Benefit to the warehouse SQL licence$2.0K to $3.3KNeeds your information

Code total, all savings$10.5K to $13.9K

ClaimsCareMemberData PlatformSharedUnassigned

HOW IT WORKS

Four steps.
You approve every change.

Your Azure scope

Read-only reader

Connect

Grant read-only access to the subscriptions in scope. Your administrator controls it and can revoke it at any time.

costinventorymetricstelemetry

Map

The mapping agent proposes a product, component and environment for each resource, with confidence and evidence. Unsupported mappings stay unassigned.

IN Investigator→CR Critic

Investigate

One agent finds savings, an independent critic challenges them, and code checks every range against recorded cost.

✓ Signed off by your engineer.
Approved by you.

Approve and report

A named engineer signs off. You approve the changes you want. The Excel report arrives every month.

COVERAGE

App Service + FunctionsContainer Apps + AKSVirtual Machines + Scale SetsAzure SQL + SQL on VMsAzure Cache for RedisStorageFront Door + Application GatewayApp Insights + Log AnalyticsAll subscription cost

Service familyComputeData and AIStorageNetworkMonitoringAll cost

AZURE TOOLS VS DYNAMO

Keep Advisor and Cost Management.
Add the part nobody owns.

Azure built-in tools compared with Dynamo
The question	Azure built-in toolsAdvisor · Cost Management · Monitor	✦ DynamoManaged FinOps by DynamoNet
Both do
Cost by subscription, resource group and tag	YesCost Management	YesSame billing data, amortized
Resource recommendations	YesAdvisor, per resource type	YesChecked against peaks, not averages
Budgets and anomaly alerts	YesCost alerts	PartlyCost changes explained monthly; keep Azure cost alerts
Only Dynamo
Cost by product, component and environment	PartlyOnly as good as your tags	YesAI mapping with confidence and evidence
Shared costs split by real usage	PartlyManual allocation rules	YesSplit by requests or ingested GB
Every finding challenged before you see it	NoYour team evaluates	YesIndependent AI critic, engineer sign-off
Savings ordered by risk, with evidence	PartlyFlat list	YesRisk, confidence, rollback, evidence
Someone owns the optimisation	NoSelf-service	YesManaged service, you keep approval
Monthly report your CFO can open	PartlyExports and dashboards	YesExcel workbook, every month

INSIDE THE ANALYSIS

Patterns, not averages.
Context, not guesses.

Demand by hour, cost by component, and the dependencies between them. This is what the agents look at before a recommendation reaches you.

▥

Demand profile

Claims Portal · requests by weekday and hour, UTC

Mon

Tue

Wed

Thu

Fri

Sat

Sun

0006121823

Who sends requestsProbes and bots separated

Other traffic 89%Probes 6%Bots 5%

▤

Cost per component

Top components · October 2026 · direct plus allocated shared

Claims APIClaims

$15.9K

SchedulerCare

$14.0K

IngestionData Platform

$13.8K

WarehouseData Platform

$8.6K

Claims PortalClaims

$7.8K

Member PortalMember

$7.8K

Shared cost

Front Door · $3,120
Split by request share. Logs split by ingested GB.

Claims 46%Member 38%Care 16%

See the allocation rules →

⌘

Project map with provenance

Clustered by project

Every resource, grouped by the project it serves and sized by cost. Observed traffic and configured links are different kinds of evidence. Choose a group to highlight it.

$92,910 in October 2026. Projects 85.2% · Shared 13.5% · Unassigned 1.3%

A configured permission is not proof of traffic, and traffic is not proof of ownership. Both stay visible in the mapping evidence.

TRUST AND SECURITY

Read-only.
And specific about it.

Access, data coverage and uncertainty are part of the report, not small print.

Discuss your security requirements →

ACCESS

You control the scope

Read-only roles on the subscriptions you choose. The reader has no permission to change infrastructure.

Bounded, checked, accountable

Agents work on a bounded evidence packet. An independent critic reviews every finding. Code computes costs and checks every saving.

MINIMIZATION

No secrets or payloads

No Key Vault secret values, connection strings, database contents, storage contents or request bodies.

APPROVAL

You approve every change

We prepare each change with its evidence and rollback. Nothing is applied without your team's approval.

PROCESSING

Terms before connection

Model provider, processing route, storage and retention are agreed in writing before any access.

EVIDENCE

Missing data stays missing

Gaps are listed in the report with what they block and how to close them. Nothing is guessed.

PLANS

Start with a free scan.
Hand us the optimisation.

Free scan

Start here

Free

Read-only. The front door.

For Azure estates spending 20k or more a month. We map your estate, find the first savings and walk you through the workbook.

Read-only access, revocable any time
AI project mapping of your estate
First recommendations with estimated savings
Excel report and engineer walkthrough

Get a free scan

Managed FinOps

Monthly

We own cost optimisation and reporting.

Continuous mapping, investigation and reporting. We prepare every change with its evidence and rollback; you approve it.

Monthly Excel report and walkthrough
New savings investigated every month
Mapping kept current as the estate changes
Realized savings measured after each change

Talk about managed FinOps

Deep dive

Fixed scope

One application or one cost question.

A focused investigation of demand, capacity, shared costs and commitments for a single workload or decision.

Scope agreed upfront
Workload and dependency analysis
Recommendations ordered by risk

Discuss a deep dive

How we work

We own cost optimisation and reporting. Your team keeps approval rights over every change.

A NAMED ENGINEER, NOT A BLACK BOX

Djordje Petrovic

Founder · DevOps and Site Reliability Engineer

DynamoNet is an independent Azure FinOps company in Belgrade, Serbia, working with teams across the EU and US. Every report is signed off by a named engineer.

Meet DynamoNet ↗

FAQ

Good questions.
Clear answers.

Something specific to your estate?

Ask Djordje →

What exactly does DynamoNet do?

We run FinOps for your Azure estate as a managed service. Dynamo, our product, proposes a product, component and environment for each resource, keeps unassigned cost visible, investigates savings and keeps the evidence. We own the optimisation work and the monthly reporting. Your team keeps approval rights over every change.

What does the AI actually do, and what does it not do?

One AI agent proposes mappings and investigates savings from a bounded evidence packet. A second, independent AI critic challenges each finding and rejects what the evidence does not support. Deterministic code computes costs and allocations from your billing data. The AI proposes each saving range with its arithmetic, and code checks it against the recorded cost of the cited resources and flags any range above it. A named engineer approves the report.

How does AI project mapping work?

The mapping agent reads names, tags, resource groups, deployment pipelines, Application Insights dependencies, managed identities and observed traffic. It proposes a product, component and environment for each resource, with a confidence level (high, medium or low) and the evidence behind it. An independent review checks each boundary; what it cannot support stays unassigned and visible. Owners confirm or correct proposals in the walkthrough.

Which Azure services do you cover?

App Service and Functions, Container Apps and AKS, Virtual Machines and Scale Sets, Azure SQL and SQL Server on VMs, Azure Cache for Redis, Storage, Front Door and Application Gateway, Application Insights and Log Analytics, plus all cost in your subscriptions, so nothing on the bill is left out of the report.

What access do you need? Do you change anything?

Read-only access, scoped with your Azure administrator and revocable at any time. Typical sources are Cost Management, Resource Graph, Azure Monitor and Application Insights. Nothing changes in your estate without your approval: we prepare each change with its evidence and rollback, and your team approves it.

What does the monthly report look like?

An Excel workbook (a project there is one of your products or applications): Summary, Recommendations, Cost by project, Cost by service, Monthly trend, Resources, Project mapping and Evidence gaps. Every number traces back to its source and allocation rule. You can download an example from this site.

How are savings calculated?

The AI proposes each estimated saving from your billing and usage data and shows its arithmetic. Code checks every range against the recorded cost of the cited resources before an engineer signs it off. After a change, realized savings are measured against the same usage priced at the old configuration over a full billing period, so changes in demand or prices do not distort the figure.

Where does our data go?

Before connecting, we agree the data scope, storage, retention, model provider and processing route in writing. Evidence is minimized before AI analysis. We never need database contents, storage contents, request bodies, Key Vault secrets or application-setting values.

Can you work with our MSP or CSP?

Yes. The report and recommendations can be shared with your engineering team, MSP or CSP, and we agree who approves and who implements each change.

Free Read-only scan of your Azure estate

See your Azure bill by product.
And what to cut first.

A free read-only scan for estates spending 20k or more a month: your estate mapped, the first savings found,
and the workbook walked through with an engineer.

Get a free read-only scan ↗Download the sample report (.xlsx) ↓

Azure cost, mapped to your products.Savings found,checked, delivered.

Project map

Three things.Every month.

Your resources, mapped toproduct, component, environment.

Project mapping

Claims

Care

Savings you can approve,because you can see why.

Recommendations

One Excel workbook.Every number traceable.

Four steps.You approve every change.

Connect

Map

Investigate

Approve and report

Keep Advisor and Cost Management.Add the part nobody owns.

Patterns, not averages.Context, not guesses.

Demand profile

Cost per component

Project map with provenance

Read-only.And specific about it.

You control the scope

Bounded, checked, accountable

No secrets or payloads

You approve every change

Terms before connection

Missing data stays missing

Start with a free scan.Hand us the optimisation.

Free scan

Managed FinOps

Deep dive

Djordje Petrovic

Good questions.Clear answers.

See your Azure bill by product.And what to cut first.

Azure cost, mapped
to your products.
Savings found,
checked, delivered.

Three things.
Every month.

Your resources, mapped to
product, component, environment.

Savings you can approve,
because you can see why.

One Excel workbook.
Every number traceable.

Four steps.
You approve every change.

Keep Advisor and Cost Management.
Add the part nobody owns.

Patterns, not averages.
Context, not guesses.

Read-only.
And specific about it.

Start with a free scan.
Hand us the optimisation.

Good questions.
Clear answers.

See your Azure bill by product.
And what to cut first.