What counts as an idle resource?

Anything you pay for that does no useful work: an instance with almost no CPU or network traffic for weeks, a disk attached to nothing, an unused IP address, a load balancer with no targets, or snapshots of volumes that no longer exist.

Do stopped instances still cost money?

Yes, partly. Compute charges stop, but attached disks and static or Elastic IP addresses keep billing. On Azure, the VM must be deallocated, not just shut down, to stop compute charges.

How long should something be idle before I remove it?

Long enough to cover its normal cycle. AWS Compute Optimizer and Google Cloud look at 14 days by default, and CloudYali's idle EC2 check needs 30. Monthly jobs need at least a month.

Optimization

Idle resources

Idle resources are cloud resources you still pay for that do no useful work: instances with almost no CPU or network traffic, unattached disks, unused IP addresses, and snapshots or load balancers nobody uses. They are the simplest cloud waste to remove, because stopping or deleting them usually changes nothing for users.

In plain English

Think of the gym membership you stopped using in February. Nothing breaks if you cancel it; the only cost of keeping it is the monthly fee, which is exactly why nobody notices.

Worked example

Worked example. The idle resources found in Northwind Labs' AWS accounts in March, priced per month.
3 EC2 instances under 2% CPU, with almost no network traffic, for 30 days	$1,100
A development RDS database with no connections for 30 days	$500
12 unattached EBS volumes, 5 TB in all	$400
Idle resources found in March	$2,000

Removing them, after a final snapshot of anything worth keeping, saves $2,000 a month, about 1.7% of the $115,000 bill. Stopping the instances alone saves their compute; the volumes keep billing until they are deleted.

Why idle resources matter

Idle resources are what's left behind: a test environment nobody tore down, a volume detached when its instance was terminated, an IP address kept just in case. Each is small, but they add up quietly, because nobody owns them and nothing alerts when they sit unused.

They are also the safest saving to take. Unlike rightsizing, removing something nobody uses carries no performance risk, which is why most cost clean-ups start here.

Common mistakes

Assuming stopped means free. Stopping an instance ends its compute charges, but its disks and static or Elastic IP addresses keep billing, on AWS, Azure and Google Cloud alike. An Azure VM shut down from inside the operating system stays allocated and keeps billing for compute until it is deallocated.
Deleting without checking. Low CPU isn't proof that something is unused: a standby database or a disaster-recovery server can look idle by design. Confirm with the owner, tag what must stay, and snapshot before deleting.
Cleaning up once. New idle resources appear every week. Make detection a scheduled job with an owner, not an annual tidy-up.
Ignoring the small items. Public IPv4 addresses, old snapshots and empty load balancers each cost little, but there are often hundreds of them. On AWS, a public IPv4 address costs the same whether it's in use or idle.

Where to find it in AWS, Azure and Google Cloud

Where to find Idle resources by cloud provider
Cloud	What it is called	Where to find it
AWS	Compute Optimizer idle recommendations	Compute Optimizer flags idle EC2 instances (peak CPU under 5% and network under 5 MB a day over 14 days), idle or unattached EBS volumes, NAT gateways, RDS and Aurora databases and more. Cost Optimization Hub lists them alongside rightsizing.
Azure	Azure Advisor	Advisor's shutdown recommendations flag VMs with low CPU and outbound network use. Deallocating a VM ends compute charges; its managed disks and static public IPs keep billing until deleted.
Google Cloud	Idle resource recommendations	Recommender flags idle VMs from CPU and network use over the last 14 days by default, plus idle persistent disks, images and IP addresses. A stopped VM is not charged for CPU or memory, but its disks and static IPs are.

How CloudYali helps

CloudYali checks AWS, Azure and Google Cloud every day for idle and unused resources, such as instances with almost no CPU, unattached disks and unused IP addresses, prices most of them per month, and tracks each one until it is removed.

See the feature Read how it works in the docs

Optimization
Cloud waste
Cloud waste is cloud spend that buys nothing the business needs: idle resources, over-provisioned resources and commitments paid for but not used. It's measured in money rather than resources, and removing it lowers the bill without changing what users get, which is why cost optimization usually starts there.
Optimization
Rightsizing
Rightsizing is changing a cloud resource to the smallest size or type that still meets its workload's real needs, based on measured CPU, memory, storage and network use. It's one of the largest usage savings in FinOps, and it should come before buying commitments, so you don't commit to capacity you don't need.
Allocation & accountability
Cost allocation tags
Cost allocation tags are key-value labels, such as team=checkout, attached to cloud resources so their cost can be grouped by owner, product or environment. On AWS a tag key only reaches billing data after it is activated for cost allocation, and tag keys and values are case-sensitive.
Allocation & accountability
Tag coverage
Tag coverage is the share of cloud cost that carries a given tag, such as team, over a period: tagged cost ÷ total cost. The rest is untagged cost that nobody can allocate, so tag coverage shows how much of the bill your allocation actually reaches.
Governance & planning
Cost anomaly
A cost anomaly is spend well outside what a service, account or team normally costs, such as a day costing several times its recent average. Anomaly detection compares each day with a baseline learned from history and alerts on the difference, so a runaway job is caught in days, not at the end of the month.

Frequently asked questions

What counts as an idle resource?
Anything you pay for that does no useful work: an instance with almost no CPU or network traffic for weeks, a disk attached to nothing, an unused IP address, a load balancer with no targets, or snapshots of volumes that no longer exist.
Do stopped instances still cost money?
Yes, partly. Compute charges stop, but attached disks and static or Elastic IP addresses keep billing. On Azure, the VM must be deallocated, not just shut down, to stop compute charges.
How long should something be idle before I remove it?
Long enough to cover its normal cycle. AWS Compute Optimizer and Google Cloud look at 14 days by default, and CloudYali's idle EC2 check needs 30. Monthly jobs need at least a month.