Description
Please add your +1 👍 to let us know you have encountered this
Status: RESOLVED
Overview:
A previous PR enabled the ability to configure logging for a state-machine that was managed via a custom resource. This introduced a bug where the logging resource policy became too large because of the name of the log group. This could break the deploys of existing applications that contained this custom resource.
Complete Error Message:
@aws-cdk--aws-dynamodb.ReplicaProvider/Provider/waiter-state-machine (Providerwaiterstatemachine5D4A9DF0) Resource handler returned message: "Invalid Logging Configuration: The CloudWatch Logs Resource Policy size was exceeded. We suggest prefixing your CloudWatch log group name with /aws/vendedlogs/states/. (Service: AWSStepFunctions; Status Code: 400; Error Code: InvalidLoggingConfiguration; Request ID: 4ba96f98-4be5-450c-a069-3d4cbf93271a; Proxy: null)"
Workaround:
Lock your cdk version to v1.115.0 or below.
Solution:
Revert in progress: #28699
Related Issues:
Original Issue:
Title: (custom-resources): Default logging configuration hits Cloudwatch Logs Resource Policy size limit
Describe the bug
Upgrading aws-cdk > 2.115 adds a default logging configuration to custom resources.
When using aws_dynamodb.Table with replication_regions specified this automatically generates a @aws-cdk--aws-dynamodb.ReplicaProvider.NestedStackResource which contains Providerwaiterstatemachine and post upgrade also a new log group
When deploying the following error is returned
@aws-cdk--aws-dynamodb.ReplicaProvider/Provider/waiter-state-machine (Providerwaiterstatemachine5D4A9DF0) Resource handler returned message: "Invalid Logging Configuration: The CloudWatch Logs Resource Policy size was exceeded. We suggest prefixing your CloudWatch log group name with /aws/vendedlogs/states/. (Service: AWSStepFunctions; Status Code: 400; Error Code: InvalidLoggingConfiguration; Request ID: 4ba96f98-4be5-450c-a069-3d4cbf93271a; Proxy: null)"
Expected Behavior
Default logging configuration should be configured in a way to prevent the above error
Current Behavior
Upgrading to aws-cdk >=2.116 generates a logging configuration with what I believe is no name/prefix specified to the log group
"ProviderwaiterstatemachineLogGroupDD693A98": {
"Type": "AWS::Logs::LogGroup",
"Properties": {
"RetentionInDays": 731,
},
"UpdateReplacePolicy": "Retain",
"DeletionPolicy": "Retain",
"Metadata": {
"aws:cdk:path": "my-stack/@aws-cdk--aws-dynamodb.ReplicaProvider/Provider/waiter-state-machine/LogGroup/Resource"
}
}
Given the nested stack is auto generated I see no way to override the properties here.
Reproduction Steps
Using the below construct with aws-cdk <=2.115 deploys successfully
table = aws_dynamodb.Table(
self,
"my-table",
table_name="my-table",
partition_key={"name": "pk", "type": aws_dynamodb.AttributeType.STRING},
removal_policy=RemovalPolicy.RETAIN,
replication_regions=["eu-central-1"],
billing_mode=aws_dynamodb.BillingMode.PAY_PER_REQUEST,
encryption=aws_dynamodb.TableEncryption.AWS_MANAGED,
point_in_time_recovery=True,
)
Upgrading aws-cdk > 2.115 causes the above error due to the newly added log group
Possible Solution
Applying the prefix recommended in the above error message /aws/vendedlogs/states/ similar to the recommendation [here] (https://docs.aws.amazon.com/step-functions/latest/dg/cw-logs.html) to the log group name?
Additional Information/Context
No response
CDK CLI Version
2.116.0
Framework Version
No response
Node.js Version
18
OS
Linux
Language
Python
Language Version
No response
Other information
No response