Solution · AWS cost optimization

AWS savings you can trust — and apply.

Over 30 checks across compute, databases, storage, networking, and commitments. Each finding is reviewed for safety, ranked, and comes with verified steps for your exact resource.

Book a demo The Cost product

Read-only IAM role Savings from your real bill Nothing runs automatically

Dilanix is cross-platform — AWS is the first provider it connects to, with more cloud providers coming.

Ranked, reviewed savings — open one, then How to fix for verified steps.

Workspace

Recommendations

AMAlex

Insights

Recommendations

A unified, cross-product inbox of optimization recommendations — never gated by a single product's own access grant.

5recommendations

1need your input

4ready to fix

AllActiveDismissedResolvedAll products All analyzers Show covered Search recommendations…

RDS db instanceorders-dbdb.r6g.2xlarge | eu-west-1

Up to1,840 USDper monthly

Ready to fix

Why this matters

orders-db peaked at 31% CPU and 42% of its memory over the last 14 days.

orders-db is sized for roughly twice its real load. Moving one size down within the same Graviton family keeps the engine, storage and Multi-AZ setup unchanged.

high confidence

Recommended action

low risk

Downsize to db.r6g.xlarge

Change the instance class during the next maintenance window. Storage, engine version and parameter group stay as they are.

Saves 1,840.00 USD / month

Validation

CPU stays below 70% for a week after the change.
Application error rates stay flat.

Rollback

Modify the instance back to db.r6g.2xlarge.

+ Halves the instance price
+ Same engine and storage

− A brief failover while the class changes

Evidence (4)

1CPU averaged 12% and peaked at 31% over 14 days (CloudWatch, 5-minute data).
2Freeable memory never fell below 58% of the instance's 64 GiB.
3db.r6g.xlarge costs 0.52 USD/hour less in eu-west-1 (AWS Price List).
4The instance is reached by orders-api and billing-worker only (VPC flow logs).

aws.rds.rightsizing · first detected Oct 9, 08:00 AM · last seen Oct 17, 07:58 AM

Before you continue

2 items to check before this can be applied.

Confirm no batch job is scheduled in the maintenance window.
Check that the latest automated snapshot completed.

Risks

The class change restarts the instance; Multi-AZ shortens it to a failover.

Environmentstaging14 resources | eu-west-1

Up to1,120 USDper monthly

Needs your input

Savings PlanCompute · 3.10 USD/hourEC2 and Fargate

Up to612.40 USDper monthly

Ready to fix

EC2 volumeapi-data-volvol-0c2d4e6f8a1b3c5d7 | eu-west-1

Up to45.06 USDper monthly

Ready to fix

EC2 nat gatewaydev-natgw-1nat-0a1b2c3d4e5f6a7b8 | eu-west-1

Up to35.04 USDper monthly

Ready to fix

How it works

From finding to fix, without the guesswork.

A recommendation only reaches you after it has been checked, reviewed, and ranked.

Detect

Every hour, each resource is checked against its bill, its CloudWatch metrics, current AWS prices, and what depends on it.

Review

Each finding is reviewed before you see it: what the resource is for, whether it is really unused, and what could break.

Rank

Monthly saving × confidence ÷ effort. The best come first, and each resource appears in only one recommendation.

Fix

Console, Terraform, and CLI steps for your resource. You run them; the recommendation resolves itself when the change lands.

What it checks

Over 30 AWS checks, grouped by where the money goes.

Rightsizing

EC2 instances and Auto Scaling groups
RDS instances and storage
ElastiCache nodes
ECS Fargate tasks and ECS capacity on EC2
Burstable instances paying for extra CPU credits
EBS volumes with more IOPS or throughput than they use

Idle and unused

Idle EC2, RDS, ElastiCache, DynamoDB, and Fargate
Load balancers and NAT gateways without traffic
Instances stopped for weeks but still paying for storage
Unattached EBS volumes and unused Elastic IPs
Environments that can sleep outside working hours

Storage

EBS gp2 to gp3
S3 Intelligent-Tiering
Old S3 object versions and incomplete uploads
ECR image lifecycle

Networking

NAT gateway consolidation
Free S3 and DynamoDB gateway endpoints instead of NAT traffic

Modernization

Previous-generation EC2 instances
Graviton for Lambda and Fargate
Fargate Spot
DynamoDB capacity mode and Lambda provisioned concurrency
RDS Multi-AZ standbys you may not need

Commitments

Compute Savings Plans over EC2 and Fargate
RDS and ElastiCache reservations
Sized only on steady usage left after the cleanup
Credit-aware — tells you when credits already cover the bill

Built on real data

Four sources behind every recommendation.

Rolled-up totals tell you spend is growing. These tell you why — and whether a change is safe.

Your real bill

The FOCUS 1.2 Data Export line by line, plus Cost Explorer totals — effective cost after discounts, credits, and commitments.

How busy it really is

CloudWatch CPU, memory, network, and connections over at least 14 days before anything is called idle or oversized.

Current AWS prices

Savings come from the AWS Price List for your region and your bill — never estimated by AI.

What depends on it

The infrastructure graph shows who talks to a resource, so a “unused” database that something still reads is never flagged.

FAQ

AWS cost optimization, answered

Is Dilanix only for AWS?

No. Dilanix is a cross-platform cost and infrastructure platform, built provider-neutral from the start. AWS is the first provider it connects to today; more cloud providers are coming.

How is this different from AWS Cost Explorer or Trusted Advisor?

They tell you something may be oversized or idle. Dilanix checks the finding against your bill, metrics, current prices, and what depends on the resource, removes the ones that are unsafe, ranks what is left by saving, confidence, and effort, and gives you verified steps to apply each one.

Does Dilanix change anything in my AWS account?

No. The IAM role it uses can only describe, list, and read. Every How to fix guide is instructions for you to follow — Dilanix never runs them.

Where do the savings numbers come from?

From current AWS prices for your region and your own billing data. AI is used to review findings and write the explanation, never to produce a dollar amount.

Are the CLI commands and Terraform safe to copy?

Every CLI command is validated against AWS's own API definitions, every Terraform resource against the provider documentation, and every resource ID must exist in your data. Anything that fails is removed and listed. Terraform is always a generic example, because Dilanix never sees your code or state.

What is FOCUS cost data and why does it matter?

FOCUS (FinOps Open Cost & Usage Specification) is a standard billing export. Its line-by-line rows carry billed, effective, list, and contracted cost per resource — the detail needed to know what a single resource really costs.

Does it work with several AWS accounts?

Yes. One connection can hold several accounts, and an AWS Organization's management account brings the bill for every member account.

Find out what your AWS really needs to cost.

Give Dilanix read-only access, let the first sync finish, and walk through your ranked savings with us.

Book a demo

AWS savings you can trust — and apply.

Over 30 checks across compute, databases, storage, networking, and commitments. Each finding is reviewed for safety, ranked, and comes with verified steps for your exact resource.

Read-only IAM role Savings from your real bill Nothing runs automatically

Dilanix is cross-platform — AWS is the first provider it connects to, with more cloud providers coming.