Amazon SQS integrates with the AWS Key Management Service (KMS) to manage KMS keys for server-side encryption (SSE). See Encryption at rest in Amazon SQS for SSE information and key management definitions. Amazon SQS uses KMS keys to validate and secure the data keys that encrypt and decrypt the messages. The following sections provide information about working with KMS keys and data keys in the Amazon SQS service.
Configuring AWS KMS permissionsEvery KMS key must have a key policy. Note that you cannot modify the key policy of an AWS managed KMS key for Amazon SQS. The policy for this KMS key includes permissions for all principals in the account (that are authorized to use Amazon SQS) to use encrypted queues.
Amazon SQS distinguishes between callers based on their AWS credentials, whether they are using different AWS accounts, IAM users, or IAM roles. Additionally, the same IAM role with different scoping policies will be treated as distinct requesters. However, when using IAM role sessions, varying only the RoleSessionName
while keeping the same IAM role and scoping policies will not create distinct requesters. Therefore, when specifying AWS KMS key policy principals, avoid relying on RoleSessionName
differences alone, as these sessions will be treated as the same requester.
Alternatively, you can specify the required permissions in an IAM policy assigned to the principals that produce and consume encrypted messages. For more information, see Using IAM Policies with AWS KMS in the AWS Key Management Service Developer Guide.
NoteWhile you can configure global permissions to send to and receive from Amazon SQS, AWS KMS requires explicitly naming the full ARN of KMS keys in specific regions in the Resource
section of an IAM policy.
Several AWS services act as event sources that can send events to Amazon SQS queues. To allow these event sources to work with encrypted queues, you must create a customer managed KMS key and add permissions in the key policy for the service to use the required AWS KMS API methods. Perform the following steps to configure the permissions.
WarningWhen changing the KMS key for encrypting your Amazon SQS messages, be aware that existing messages encrypted with the old KMS key will remain encrypted with that key. To decrypt these messages, you must retain the old KMS key and ensure that its key policy grants Amazon SQS the permissions for kms:Decrypt
and kms:GenerateDataKey
. After updating to a new KMS key for encrypting new messages, ensure all existing messages encrypted with the old KMS key are processed and removed from the queue before deleting or disabling the old KMS key.
Create a customer managed KMS key. For more information, see Creating Keys in the AWS Key Management Service Developer Guide.
To allow the AWS service event source to use the kms:Decrypt
and kms:GenerateDataKey
API methods, add the following statement to the KMS key policy.
{
"Version": "2012-10-17",
"Statement": [{
"Effect": "Allow",
"Principal": {
"Service": "service
.amazonaws.com"
},
"Action": [
"kms:Decrypt",
"kms:GenerateDataKey"
],
"Resource": "*"
}]
}
Replace "service" in the above example with the Service name of the event source. Event sources include the following services.
Configure an existing SSE queue using the ARN of your KMS key.
Provide the ARN of the encrypted queue to the event source.
When the data key reuse period expires, the producer's next call to SendMessage
or SendMessageBatch
also triggers calls to kms:Decrypt
and kms:GenerateDataKey
. The call to kms:Decrypt
is to verify the integrity of the new data key before using it. Therefore, the producer must have the kms:Decrypt
and kms:GenerateDataKey
permissions for the KMS key.
Add the following statement to the IAM policy of the producer. Remember to use the correct ARN values for the key resource and the queue resource.
{
"Version": "2012-10-17",
"Statement": [
{
"Effect": "Allow",
"Action": [
"kms:Decrypt",
"kms:GenerateDataKey"
],
"Resource": "arn:aws:kms:us-east-2:123456789012:key/
"
},
{
"Effect": "Allow",
"Action": [
"sqs:SendMessage"
],
"Resource": "arn:aws:sqs:*:123456789012:MyQueue
"
}
]
}
When the data key reuse period expires, the consumer's next call to ReceiveMessage
also triggers a call to kms:Decrypt
, to verify the integrity of the new data key before using it. Therefore, the consumer must have the kms:Decrypt
permission for any KMS key that is used to encrypt the messages in the specified queue. If the queue acts as a dead-letter queue, the consumer must also have the kms:Decrypt
permission for any KMS key that is used to encrypt the messages in the source queue. Add the following statement to the IAM policy of the consumer. Remember to use the correct ARN values for the key resource and the queue resource.
{
"Version": "2012-10-17",
"Statement": [
{
"Effect": "Allow",
"Action": [
"kms:Decrypt"
],
"Resource": "arn:aws:kms:us-east-2:123456789012:key/
"
},
{
"Effect": "Allow",
"Action": [
"sqs:ReceiveMessage"
],
"Resource": "arn:aws:sqs:*:123456789012:MyQueue
"
}
]
}
When the principal in a key policy statement is an AWS service principal, you can use the aws:SourceArn
or aws:SourceAccount
global condition keys to protect against the confused deputy scenario. To use these condition keys, set the value to the Amazon Resource Name (ARN) of the resource that is being encrypted. If you don't know the ARN of the resource, use aws:SourceAccount
instead.
In this KMS key policy, a specific resource from service that is owned by account 111122223333
is allowed to call KMS for Decrypt
and GenerateDataKey
actions, which occur during SSE usage of Amazon SQS.
{
"Version": "2012-10-17",
"Statement": [
{
"Effect": "Allow",
"Principal": {
"Service": "<replaceable>service</replaceable>.amazonaws.com
"
},
"Action": [
"kms:GenerateDataKey",
"kms:Decrypt"
],
"Resource": "*",
"Condition": {
"ArnEquals": {
"aws:SourceArn": [
"arn:aws:<replaceable>service<::111122223333:resource
"
]
}
}
}
]
}
When using SSE enabled Amazon SQS queues, the following services support aws:SourceArn
:
Amazon SNS
Amazon S3
CloudWatch Events
AWS Lambda
CodeBuild
Amazon Connect Customer Profiles
AWS Auto Scaling
Amazon Chime
The data key reuse period defines the maximum duration for Amazon SQS to reuse the same data key. When the data key reuse period ends, Amazon SQS generates a new data key. Note the following guidelines about the reuse period.
A shorter reuse period provides better security but results in more calls to AWS KMS, which might incur charges beyond the Free Tier.
Although the data key is cached separately for encryption and for decryption, the reuse period applies to both copies of the data key.
When the data key reuse period ends, the next call to SendMessage
or SendMessageBatch
typically triggers a call to the AWS KMS GenerateDataKey
method to get a new data key. Also, the next calls to SendMessage
and ReceiveMessage
will each trigger a call to AWS KMS Decrypt
to verify the integrity of the data key before using it.
Principals (AWS accounts or users) don't share data keys (messages sent by unique principals always get unique data keys). Therefore, the volume of calls to AWS KMS is a multiple of the number of unique principals in use during the data key reuse period.
To predict costs and better understand your AWS bill, you might want to know how often Amazon SQS uses your KMS key.
NoteAlthough the following formula can give you a very good idea of expected costs, actual costs might be higher because of the distributed nature of Amazon SQS.
To calculate the number of API requests (R
) per queue, use the following formula:
R = (B / D) * (2 * P + C)
B
is the billing period (in seconds).
D
is the data key reuse period (in seconds).
P
is the number of producing principals that send to the Amazon SQS queue.
C
is the number of consuming principals that receive from the Amazon SQS queue.
In general, producing principals incur double the cost of consuming principals. For more information, see Understanding the data key reuse period.
If the producer and consumer have different users, the cost increases.
The following are example calculations. For exact pricing information, see AWS Key Management Service Pricing.
Example 1: Calculating the number of AWS KMS API calls for 2 principals and 1 queueThis example assumes the following:
The billing period is January 1-31 (2,678,400 seconds).
The data key reuse period is set to 5 minutes (300 seconds).
There is 1 queue.
There is 1 producing principal and 1 consuming principal.
(2,678,400 / 300) * (2 * 1 + 1) = 26,784
Example 2: Calculating the number of AWS KMS API calls for multiple producers and consumers and 2 queues
This example assumes the following:
The billing period is February 1-28 (2,419,200 seconds).
The data key reuse period is set to 24 hours (86,400 seconds).
There are 2 queues.
The first queue has 3 producing principals and 1 consuming principal.
The second queue has 5 producing principals and 2 consuming principals.
(2,419,200 / 86,400 * (2 * 3 + 1)) + (2,419,200 / 86,400 * (2 * 5 + 2)) = 532
AWS KMS errors
When you work with Amazon SQS and AWS KMS, you might encounter errors. The following references describe the errors and possible troubleshooting solutions.
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4