Skip to content

ReservedInstancesLimitExceeded when purchasing g5.2xlarge Standard RI despite quotas showing headroom

-1

ReservedInstancesLimitExceeded when purchasing g5.2xlarge Standard RI despite quotas showing headroom (eu-north-1)

Tags: amazon-ec2, reserved-instances, aws-support, service-quotas, gpu


I'm trying to purchase 1x g5.2xlarge Standard Reserved Instance, 3-year term, No Upfront, in eu-north-1 (Stockholm), and consistently get the following error:

Error: Your current quota does not allow you to purchase the required number of Reserved Instances.
(Status Code: 400; Error Code: ReservedInstancesLimitExceeded; Request ID: 5be28959-2621-4b73-bbab-57ad4e7fc242)

Account context

  • Currently running 1x g5.2xlarge and 1x g6.xlarge on-demand (plus non-GPU instances: c6a.large, m6a.xlarge)
  • 0–2 active Reserved Instances total (none for g5.2xlarge)
  • This is the first g5.2xlarge RI purchase attempt on the account

Quotas I've already checked and ruled out

(so we don't go in circles)

  • "New Reserved Instances per month" (Service Quotas → EC2): Applied quota 20, AWS default 20 — nowhere near consumed.
  • "Running On-Demand G and VT instances" (quota code L-DB2E81BA, vCPU-based): Applied quota 32 vCPUs, current utilization 12 vCPUs (4 from g6.xlarge + 8 from g5.2xlarge) — 20 vCPUs of headroom, well above the 8 vCPUs a g5.2xlarge RI would need.

Neither of these self-service quotas explains the block. I've confirmed the purchase is for a regional RI (not zonal), so it shouldn't be constrained by AZ capacity reservation limits either.

What I suspect

This may be an account-level or region-level restriction specific to GPU-family Reserved Instances (possibly tied to billing-risk controls AWS applies to No Upfront RIs on lightly-used accounts, or GPU capacity/resale controls in eu-north-1), which doesn't appear to be exposed as an adjustable quota in the Service Quotas console at all.

AWS Support case open

*************** (Basic support plan — currently unassigned, only received an automated triage response referencing the two quotas above, which I'd already checked).


Question: Has anyone hit this specifically for GPU instance families (g4/g5/g6/p-series) when purchasing RIs, and if so — what actually resolved it? Was it a manual account review, a specific quota name I'm missing, or a Support escalation to a particular team?

*Edit: Removed case ID - Roman Z.

2 Answers
0

The error you're encountering indicates that your account has reached a limit for purchasing Reserved Instances, but as you've correctly identified, the standard self-service quotas don't explain the block. Based on the error code 'ReservedInstancesLimitExceeded,' your current quota does not allow you to purchase the required number of Reserved Instances, even though your visible quotas show available capacity.

While I cannot see the specific internal restriction affecting your account, there are a few possibilities for what might be causing this:

First, there may be account-level purchasing restrictions that aren't exposed in the Service Quotas console. These can sometimes apply to newer accounts, accounts with limited payment history, or when purchasing No Upfront Reserved Instances, which carry higher financial risk for AWS since there's no upfront payment.

Second, GPU instance families may have additional purchasing controls that differ from general-purpose instances. These specialized instance types can have separate reservation limits that aren't always visible in the standard quota displays.

Third, regional availability can play a role. Certain instance families may have restricted Reserved Instance availability in specific regions, particularly for newer GPU types or in regions with limited capacity.

To resolve this issue, you'll need to work directly with AWS Support through your existing case. Since you have a Basic support plan, response times may be longer, but the support team will be able to see account-level restrictions that aren't visible to you. When you add correspondence to your case in the AWS Support Center, be sure to include all the details you've provided here—the specific instance type, term length, payment option, region, and the quota checks you've already performed. This will help them investigate the root cause more quickly.

If faster response times are critical for your business needs, you might also consider reviewing your AWS Support plan options, as higher-tier plans provide more rapid responses for these types of issues.
Sources
Error codes for the Amazon EC2 API - Amazon Elastic Compute Cloud
PurchaseReservedInstancesOffering - Amazon Elastic Compute Cloud

answered a month ago

AWS
MODERATOR

reviewed a month ago

EXPERT

reviewed a month ago

0

Hi there.

I apologize for trouble you're experiencing with the error message when purchasing a Reserved Instance. I've shared your sentiments internally with our Support team.

Continue monitoring your case as correspondence will be shared once the review is complete.

— Roman Z.

AWS
MODERATOR

answered a month ago

EXPERT

reviewed a month ago

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.