New Savings Proposals: approve, test and roll back cost changesLearn more Sign in|Talk to a cloud engineer
Product tourGet a demo

Priced savings, tracked until they’re done

Rightsizing, idle resources, commitments, storage, network and Kubernetes. Each recommendation has its own dollar figure, an owner, a comment thread and a status.

Get a demoTry the interactive demo
What you're watching
  1. 1
    Select three recommendations

    The checkout-api node group rightsizing, search-indexer to Graviton and idle analytics-sandbox instances. A selection bar appears.

  2. 2
    Move them to In Progress

    Change status updates all three at once. Not Started drops from 86 to 83, and owners are notified.

  3. 3
    Add one comment to all three

    “Owners confirmed, rolling out in Saturday's window” is posted to each recommendation's thread.

Who does thisFinOps lead running the weekly savings reviewWhat you getThree priced recommendations moved forward in one pass, each with the same note on its thread.
Potential$23,480/mopriced on the resources named
Open142recommendations across 12 categories
Detectors40+rules for AWS and Azure
Statuses5from Not Started to No Longer Detected

Sample data from a fictional company.

A recommendation nobody owns never ships

A list of savings is easy to produce. Tracking each one to done is where teams lose them.

Delivery

Savings lists rot in spreadsheets

What usually happens: Findings get exported to a sheet each quarter, owners are guessed, and by the next review half the rows are stale.

How CloudLens resolves it: Each recommendation keeps its owner, comments, Jira issue and status on the recommendation itself, updated on every scan.

Resolved
Accuracy

Flat-percentage estimates

What usually happens: “Rightsizing saves 30%” makes a good slide and a bad budget.

How CloudLens resolves it: Savings are calculated from your cost and usage data and the provider’s price for that resource in that region.

Resolved
Reporting

Counting wins that aren’t real

What usually happens: A terminated instance or a lost permission looks exactly like a fix, and the savings report goes up anyway.

How CloudLens resolves it: Addressed is kept apart from realized savings, and anything that simply disappeared goes to No Longer Detected.

Resolved

How Ana runs the Monday savings review

Forty minutes that leave every priced recommendation with an owner, a ticket or a reason.

ARAna RuizFinOps lead, Lumora Retail

Lumora Retail is a fictional company. The people, names and numbers are sample data.

    1
    Mon 10:00

    Filters to Not Started, sorted by savings

    142 open recommendations worth $23,480/mo. Near the top: rightsizing the checkout-api node group, priced at $3,140/mo.
    2
    Mon 10:10

    Assigns owners in bulk

    She selects the eight EBS and snapshot items, sets the owner to Data Platform and adds one comment to all of them from the selection bar.
    3
    Mon 10:20In Progress

    Creates FINOPS-218 for checkout-api

    The Jira issue is prefilled with the resource, the change and the saving. The recommendation moves to In Progress.
    4
    Mon 10:30

    Posts the NAT gateway item to Teams

    The VPC endpoint recommendation goes to the Shared Services channel in Microsoft Teams, where the network team will see it.
    5
    Mon 10:40Ignored

    Ignores one, with a reason

    The orders-db snapshot cleanup is set to Ignored. Those snapshots are kept for an audit, and the reason stays on the thread.
    6
    Next Mon

    Checks the realized savings ledger

    The gp2 to gp3 migration on mobile-api-asg shows as realized once the bill confirmed it. Two items whose resources vanished sit in No Longer Detected and are not counted.
Sitting on a savings list nobody works?

See your recommendations, priced per resource

Get a demo

From detection to done, in one place

Owners, comments, tickets and channel posts live on the recommendation, not in a separate tracker.

Assign it, discuss it, ticket it

Jira integration
Open a recommendation to see its saving and confidence, assign an owner and keep the conversation on the thread. Creating a Jira issue moves it to In Progress. Posting it to Slack or Microsoft Teams puts it in front of the team that owns the resource.
  • Owner per recommendation
  • Comment threads
  • Jira issue moves it to In Progress
  • Post to Slack or Teams
What you're watching
  1. 1
    Assign Priya as owner

    Rightsizing the checkout-api node group to m7g.xlarge saves $3,140/mo at 91% confidence. Owner goes from Unassigned to Priya, Checkout lead.

  2. 2
    Agree on the thread

    Ana's note says p95 CPU is 18% and memory 41% over 30 days. Priya replies she will roll it out Saturday.

  3. 3
    Create FINOPS-218

    The Jira issue is linked and the recommendation moves to In Progress. Jira status syncs back.

  4. 4
    Post it to Slack

    It lands in #cloud-cost-alerts with the saving and a link back to the recommendation.

Who does thisFinOps lead and the owning engineering leadWhat you getA $3,140/mo change with an owner, a Jira issue and a rollout date.

See which detectors save the most

Insights groups savings by detector, so you can see which patterns cost the most across your estate. The Rule Catalog lists every rule with its category and threshold. Change a threshold and it applies on the next scan.
  • Insights by detector
  • Rule Catalog
  • Editable thresholds
  • Confidence score per finding
What you're watching
  1. 1
    Rank savings by detector

    Compute Savings Plan opportunity leads at $5,820/mo. Idle EC2 is next at $4,210 across 18 recommendations.

  2. 2
    Open the Rule Catalog

    Every rule shows its category, threshold and whether it is on. Lambda over-provisioned memory is switched off.

  3. 3
    Tighten the Idle EC2 rule

    The threshold changes from p95 CPU under 5% over 30 days to 14 days. It re-evaluates on the next scan.

Who does thisFinOps lead, with the platform engineer who agrees the thresholdsWhat you getDetectors tuned to how your workloads actually run, applied on the next scan.

Addressed is not the same as realized

Fix the easy ones in minutes
Marking a recommendation Addressed tells your team it is done. It only enters the realized savings ledger when the bill shows the drop. A resource that disappeared goes to No Longer Detected instead of your savings total.
  • Realized savings ledger
  • Confirmed against billing
  • Disappeared resources kept apart
What you're watching
  1. 1
    Move gp2 to gp3 through the statuses

    The mobile-api-asg migration goes from Not Started to In Progress, then Addressed. The ledger does not move yet.

  2. 2
    Wait for the bill

    Addressed means the team says it is done. Realized savings still reads $38,210 for the quarter.

  3. 3
    The ledger confirms $516/mo

    When billing shows the drop, the saving is added and this week's bar grows. No Longer Detected is never counted.

Who does thisFinOps lead reporting savings to financeWhat you getA realized savings number finance can trace back to the bill.

Five statuses, each with one meaning

Not Started

Detected and priced, waiting for someone to pick it up.

In Progress

Someone owns it, often because a Jira issue was created.

Addressed

Your team marked it done. Not yet counted as realized savings.

Ignored

Skipped on purpose, with the reason kept in the thread.

No Longer Detected

The resource is gone. Tracked separately and never counted as a win.

Why No Longer Detected is its own status

An instance terminated by an autoscaler, or a scan that lost a permission, looks the same as a fix from the outside. Counting it would inflate your savings, so CloudLens keeps it separate.

Confidence on every recommendation

Each one shows how sure the detector is, based on utilization history and pricing data, so you can start with the safest.

40+ rules in 12 categories

Compute, storage, network, load balancers, RDS, Lambda, ECS, EKS, ElastiCache, EMR, Redshift and serverless.

AWS and Azure

The same statuses, owners and threads for both clouds, priced from each provider’s own billing data.

The rest of the savings workflow

You are here

Recommendations

Priced per resource, owned and tracked.

02 · Fix fast

Quick Wins

Safe fixes with the command ready.

03 · Change safely

Savings Proposals

Approve, dry-run, apply, roll back.

04 · Automate

Scheduler

Stop non-prod when nobody's using it.

Recommendation questions

From your own cost and usage data and the provider’s pricing for that region and resource, not a flat percentage. Where a figure is an estimate, the recommendation says so.

Yes. Select rows and use the selection bar to change their status or add the same comment to all of them.

No. It creates the issue, links it and moves the recommendation to In Progress. Changes happen through Savings Proposals or your own process.

The recommendation moves to No Longer Detected. It is not counted as addressed or as realized savings.

Ready to see CloudLens in action?

Connect a read-only AWS role or Azure service principal. We'll walk you through your bill, your security graph and the first things worth fixing.