SAM Doctor

ERROR REFERENCE

CloudFormation was temporarily unavailable

AWS rejected the CloudFormation API call because the control plane was unavailable. This response is a service interruption, not proof that your template or IAM policy is wrong.

WHAT IT MEANS

What this error means

An error occurred (ServiceNotAvailable) when calling the CreateStack operation

CloudFormation returned a temporary 503-style service-unavailable response before the stack operation could complete. The request may still have reached AWS, so a retry should preserve the operation's identity instead of blindly creating a second stack.

FIX

How to fix it

  1. Check the regional AWS status and the CloudFormation events for the stack before changing the template. A service interruption can resolve without any code change.
  2. Wait briefly and retry with backoff. For a new stack, provide a stable client request token so CloudFormation can distinguish a retry from a second create: 
    aws cloudformation create-stack --stack-name <stack> --client-request-token <stable-token> ...
  3. Preserve the evidence if it repeats. Record the Region, UTC timestamp, operation, request ID, and exact error before opening an AWS support case. Do not rewrite the template based on a service-availability response alone.
  4. Re-run the deployment after the service recovers. See AWS's CloudFormation troubleshooting guide for service-interruption guidance.

AUTOMATE THE TRIAGE

Diagnose this automatically

SAM Doctor recognizes the CloudFormation service interruption (high confidence), keeps it separate from throttling and generic resource failures, and points at a safe retry. Runs locally; no AWS access, no log upload. 

python -m pip install sam-doctor
sam-doctor diagnose deployment.log --format markdown

RELATED