AWS service quotas: check your usage before you hit one
The Service Quotas console lists hundreds of quotas, one service at a time, and shows usage only where AWS publishes a usage metric. CloudConsole’s Service quotas checks the 16 that cause outages when they are reached, reads current usage from CloudWatch AWS/Usage metrics (or counts the resources itself where AWS publishes no metric), and flags any at 80% or more, with a Request increase button.


What it checks
Open Services → Tools → Service quotas. It checks the current region’s applied value (your limit, including increases you’ve been granted) for each of these:
| Quota | What fails when it’s reached |
|---|---|
| On-Demand vCPUs: standard (A, C, D, H, I, M, R, T, Z) | New instances and Auto Scaling launches (VcpuLimitExceeded) |
| On-Demand vCPUs: G and VT, P, and X | GPU and memory-optimised instances won’t launch |
| Spot vCPUs: standard | Spot requests stay unfulfilled |
| Elastic IP addresses * | AllocateAddress fails; new NAT gateways can’t get an IP |
| VPCs per Region * | CreateVpc fails |
| NAT gateways per Availability Zone * | CreateNatGateway fails in a full AZ (the busiest AZ is shown) |
| Network interfaces per Region | VPC Lambda functions, Fargate tasks and instances can’t get an interface |
| Lambda concurrent executions | Every function in the region is throttled (429) |
| Fargate On-Demand and Spot vCPUs | ECS and EKS tasks on Fargate won’t start |
| EBS gp3 and gp2 storage (TiB) | New volumes and snapshot restores fail |
| Application Load Balancers per Region * | CreateLoadBalancer fails |
| CloudFormation stacks | CreateStack fails for new stacks |
Usage comes from the CloudWatch AWS/Usage metric that Service Quotas names for the quota, where AWS publishes one. For the quotas marked *, AWS publishes none, so the app counts the resources itself and labels the figure “counted”.
Usage from AWS/Usage is the highest value in the last hour, at 5-minute resolution, which is what matters for “can I launch now”. Rows are sorted fullest first, and quotas at 80% or more are flagged. Quotas with a limit of 0 (often GPU vCPUs on new accounts) are listed separately as ones to request before you need them.
Request increase sends a Service Quotas increase request for the value you enter (a Pro change, with your confirmation). Increases take from minutes to days, and some are reviewed by AWS Support, so ask before the launch that needs them.
How it works, and what it can read
One ListServiceQuotas per service code and one GetMetricData request for all the usage metrics, each time it loads. AWS bills GetMetricData beyond its free tier; a load reads a few metrics for an hour, a fraction of a cent.
| What it reads | IAM actions |
|---|---|
| Applied quota values and their usage metrics | servicequotas:ListServiceQuotas, servicequotas:GetServiceQuota |
| Usage | cloudwatch:GetMetricData |
| Counted quotas | ec2:DescribeVpcs, ec2:DescribeAddresses, ec2:DescribeNatGateways, ec2:DescribeSubnets, elasticloadbalancing:DescribeLoadBalancers |
| Request increase (Pro, a change) | servicequotas:RequestServiceQuotaIncrease |
All of these reads are in the AWS managed ReadOnlyAccess policy. ViewOnlyAccess leaves some out; Recommended permissions lists the extras. If one call is denied, the rest of the tool still works and says which permission is missing.
Nothing leaves your phone. The app signs each request with your own credentials, which stay in the iOS Keychain or Android Keystore, and sends it straight to AWS. There is no CloudConcur server in between, so we never see your resources or the results.
Free to look, Pro to fix. Every check only reads, so it works on the free plan (one account) and in the built-in demo. Where the tool offers a fix, it is the app’s usual action: it needs Pro, goes through the confirmation you set for that account and is recorded in the activity log on your phone.
Do it without the app
Read one quota’s applied value and the metric AWS uses for its usage:
aws service-quotas get-service-quota --service-code ec2 --quota-code L-1216C47A \
--query 'Quota.[QuotaName,Value,UsageMetric]'
Then read that metric. For standard On-Demand vCPUs it is ResourceCount in AWS/Usage:
aws cloudwatch get-metric-statistics --namespace AWS/Usage --metric-name ResourceCount \
--dimensions Name=Service,Value=EC2 Name=Type,Value=Resource \
Name=Resource,Value=vCPU Name=Class,Value=Standard/OnDemand \
--statistics Maximum --period 300 \
--start-time 2026-11-17T09:00:00Z --end-time 2026-11-17T10:00:00Z
Quotas without a usage metric, such as VPCs or Elastic IPs, need a count from the matching describe call. To ask for more:
aws service-quotas request-service-quota-increase \
--service-code lambda --quota-code L-B99A9384 --desired-value 2000
The Service Quotas console can also create a CloudWatch alarm on a quota’s usage, which is the right long-term guard for the one or two that keep biting.
Questions
What is the difference between a service quota and a service limit?
Nothing: AWS renamed service limits to service quotas. Most are per account per region and adjustable; some are fixed.
Why is a quota’s value higher than the AWS default?
The tool shows the applied value, which includes increases granted to your account. The default is what a new account starts with.
Why does it say “counted”?
AWS publishes no usage metric for some quotas, such as VPCs, Elastic IPs, NAT gateways per AZ and Application Load Balancers, so the app counts them itself with the same read calls as their own screens.
Try it on your own account
CloudConsole is on the App Store for iPhone and Google Play for Android. Every tool also works in the built-in demo, if you’d rather look around before connecting an account.
Details: Tools for your AWS account in the help centre.