Skip to main content
A network module fails to deploy with:
or the equivalent for internet gateways:

​
Why this happens

AWS accounts start with a quota of 5 VPCs per region, and the same number of internet gateways. Each rvn-aws-network module instance creates one of each. Five environments — or three environments plus a couple of VPCs left over from earlier experiments — hit the ceiling.

​
Fix: share one network across environments

You rarely need a VPC per environment. Create one rvn-aws-network instance for your non-production environments and have staging, preview, and development services reference it. Keep production in its own VPC. Every module that takes a network dependency can also point at existing infrastructure by supplying the VPC and subnet IDs directly instead of a moduleGivenIdRef, so services can join a VPC that Ravion did not create. See Import into a standard module if you want to bring an existing VPC under Ravion management.

​
Fix: delete unused VPCs

Check the VPC console for VPCs from deleted projects, old tooling, or the default VPC in regions you do not use. Deleting the default VPC is safe if nothing runs in it; you can recreate it from the console later.

​
Fix: request a quota increase

VPC quotas are adjustable. Request an increase in Service Quotas — make sure the console is set to the region you deploy to: The two quotas must match, so request both. Increases up to a few dozen are usually approved automatically; larger ones go to support and can take a day.

​
Fix: use another region

Quotas are per region. If you need an environment right now and the account is full in one region, deploying to a neighbouring region works — but keep latency to your database in mind.