AWS ACTION : aws-cloudformation-github-deploy action running indefinitely even after the stack gets completely created

[arqgroup-snehalthakre]

Select Topic Area

Bug

Body

I am working on creating an RDS AURORA Serverless cluster with 2 reader instances. The AWS action (aws-cloudformation-github-deploy) runs fine with 1 reader instance. Still, as soon as the second reader instance is added, the action keeps running indefinitely till it reaches the runner's maximum execution time of 360 minutes. The creation of a second reader instance adds time to stack creation but the stack creation does get complete in 24 min.

My observation when I tried to use the older version of [email protected], the stack was in CREATE_IN_PROGRESS but the aws action threw an error of stack not in CREATE_COMPLETE state and failed. Although the stack continued to be completed.

My code:
name: Deploy RDS AURORA serverless cluster on ${{ inputs.env }}
id: deploy_cluster
uses: aws-actions/aws-cloudformation-github-deploy@v1
with:
name: ${{ inputs.appName }}-rds-cluster-${{ inputs.env }}
template: .github/infrastructure/rds/rds-cluster.yaml
no-fail-on-empty-changeset: true
capabilities: CAPABILITY_NAMED_IAM, CAPABILITY_AUTO_EXPAND
parameter-overrides: "AppName=${{ inputs.appName }},Environment=${{ inputs.env }},VpcId=vpc-01fa9a364bdb89d67,Subnet1=subnet-0ecbb3f7b1fd26be7,Subnet2=subnet-053216efcb229544d,MinCapacity=1,MaxCapacity=4,DBName=${{ inputs.dbName }}"
timeout-in-minutes: 60

Stack yaml:
` TDBCluster:
Type: AWS::RDS::DBCluster
Properties:
DBClusterIdentifier: !Sub ${Environment}-test-db-cluster
Engine: aurora-postgresql
EngineMode: provisioned
EngineVersion: '15.4'
DatabaseName: !Sub ${DBName}
MasterUsername: !Sub '{{resolve:secretsmanager:${Environment}/master/rds:SecretString:username}}'
MasterUserPassword: !Sub '{{resolve:secretsmanager:${Environment}/master/rds:SecretString:password}}'
DBSubnetGroupName: !Ref TDBSubnetGroup
EnableHttpEndpoint: true
DeletionProtection: false # TODO: True for higher environment
VpcSecurityGroupIds:
- !Ref TDBSecurityGroup
ServerlessV2ScalingConfiguration:
MinCapacity: !Ref MinCapacity
MaxCapacity: !Ref MaxCapacity
DBClusterParameterGroupName: !Ref TDBClusterParamGrp
StorageEncrypted: true
BackupRetentionPeriod: 7
PreferredBackupWindow: 18:56-19:26
PreferredMaintenanceWindow: mon:12:11-mon:12:41
Port: 5432
CopyTagsToSnapshot: true
StorageType: aurora
NetworkType: IPV4
EnableCloudwatchLogsExports:
- postgresql
AvailabilityZones:
- ap-southeast-2b
- ap-southeast-2a
AutoMinorVersionUpgrade: true
EnableIAMDatabaseAuthentication: true
Tags:
- Key: Environment
Value: !Ref Environment
- Key: App Name
Value: !Ref AppName

Reader node instance for RDS Aurora Cluster

TInstance2:
Type: AWS::RDS::DBInstance
DependsOn:
- TDBCluster
- TInstance1
Properties:
DBInstanceIdentifier: !Sub "${Environment}-test-db-reader-2"
DBClusterIdentifier: !Ref TDBCluster
DBParameterGroupName: !Ref DBParameterGroup
DBInstanceClass: db.serverless
Engine: aurora-postgresql
PubliclyAccessible: false
AutoMinorVersionUpgrade: true
StorageEncrypted: true
Tags:
- Key: Environment
Value: !Ref Environment
- Key: AppName
Value: !Ref AppName`

[TucanoWeb]

Please, post your yaml

[arqgroup-snehalthakre]

@TucanoWeb - Updated the description. Please check.

[TucanoWeb]

Tks.

Chnage to v2 version and set timeout considering the runtime observed + buffer.

uses: aws-actions/aws-cloudformation-github-deploy@v2
timeout-in-minutes: 30

Also, Dependsn can be removed, because the new version manages instance dependencies automatically.

• Add an explicit waiter to workflow:

- name: Wait for stack completion
  run: |
    aws cloudformation wait stack-create-complete \
      --stack-name ${{ inputs.appName }}-rds-cluster-${{ inputs.env }}

[arqgroup-snehalthakre]

There is no v2 version. The latest available version is v1.3.0

[TucanoWeb]

Yes, my bad. Do you tried with latest version (1.3)?

[arqgroup-snehalthakre]

Yes, I tried, the same issue occurred.