aws-cloudformation / aws-cloudformation/cloudformation-coverage-roadmap
AWS::EC2::InstanceConnectEndpoint DELETE_FAILED
- Dominant language
- No language data
- Stars
- 1.1k
- Forks
- 62
- PR merge metrics
- No merged PRs in 30d
Description
### Name of the resource
AWS::EC2::InstanceConnectEndpoint
### Issue Description
The `AWS::EC2::InstanceConnectEndpoint` CloudFormation Construct can not successfully be deleted unless it has already completed deployment. In other words, issuing a DELETE of the resource while it is still in the "provisioning" state will fail.
The construct needs to be allowed to progress to a state (e.g. reach CREATE_COMPLETE) where it will accept a DELETE prior to having the DELETE issued to it.
This issue is 100% reproducible, and mostly shows up in rollbacks.
### Expected Behavior
Delete of the `AWS::EC2::InstanceConnectEndpoint` should succeed in the case of needing to delete before provisioning has completed.
### Observed Behavior
During rollbacks the `AWS::EC2::InstanceConnectEndpoint` will often fail to delete because it is still provisioning. If, after failing rollback, I wait until provisioning is complete and then issue `destroy` for the stack, it will succeed.
### Test Cases
Here is some CDK code:
```js
const vpc = new ec2.Vpc(this, 'my-vpc', {
vpcName: 'my-vpc',
availabilityZones: ['us-east-1a', 'us-east-1b'],
natGateways: 0,
ipAddresses: ec2.IpAddresses.cidr('172.31.0.0/16'),
});
const eiceSubnet = vpc.isolatedSubnets[0];
new ec2.CfnInstanceConnectEndpoint(this, 'my-instance-connect-endpoint', {
subnetId: eiceSubnet.subnetId,
securityGroupIds: [instanceConnectSecurityGroup.securityGroupId],
});
// TODO: insert some resource that fails at runtime.
// e.g. Lambda-backed CustomResource which is just `throw new Error()`
```
### Other Details
_No response_
Contributor guide
Assessment
This issue has not been assessed yet.