The AWS Service Quotas That Will Take Down Your Production at 3 AM (And You Cannot Raise Them Fast Enough) | by Illya Yalovoy | May, 2026 | MediumSitemapOpen in appSign up<br>Sign in

Medium Logo

Get app<br>Write

Search

Sign up<br>Sign in

Member-only story

The AWS Service Quotas That Will Take Down Your Production at 3 AM (And You Cannot Raise Them Fast Enough)

Illya Yalovoy

13 min read·<br>Just now

Listen

Share

Hard limits, scaling lags, and the architectural walls that no support ticket can fix.<br>Your pager fires at 3:12 AM. Users are getting 503s. You check the dashboard — Lambda is throttling at 1,000 concurrent executions. You open a quota increase request and see the estimated response time: 1–3 business days. It is Saturday. This is the moment you learn the difference between a soft limit and a hard wall.<br>The Two Categories of AWS Limits That Matter<br>Every AWS account ships with two kinds of limits, and the difference between them is the difference between a minor inconvenience and a 3 AM architecture redesign.<br>The first kind is adjustable quotas. Lambda concurrent executions (default 1,000 per region), EC2 on-demand vCPUs per instance family, API Gateway requests per second. You can raise these through the Service Quotas console or a support ticket. This sounds reassuring until you learn the timeline: most increases take 1–3 business days. Some, like GPU instances or dedicated hosts, take weeks. If your traffic spike is happening right now, a support ticket is not a solution. It is a post-mortem action item.<br>The second kind is hard limits. NAT Gateway caps at 55,000 simultaneous connections per destination. S3 gives you 5,500 GET and 3,500 PUT requests per second per prefix. DynamoDB on-demand tables will…

Written by Illya Yalovoy<br>5 followers<br>·8 following

Help

Status

About

Careers

Press

Blog

Privacy

Rules

Terms

Text to speech