Repository containing code for a self-contained service to generate unique "references" for records transferred to The National Archives (TNA).
The service consists of three main components:
- Lambda function: the Lambda generates the "references" based on a incremented counter;
- DynamoDb Table: the table contains the counter to generate the unique "references";
- API Gateway: Provides access to the service from other AWS services and external clients outside of AWS.
The lambda contains the code for generating references.
This Lambda function is triggered by an api gateway event and context object. It retrieves the current value of the counter from DynamoDB, increments it,
and stores the updated value back in DynamoDB. It then generates a unique “reference” using the counter value. Finally, it returns the generated “references” json to the caller like so:
[ "M3H5J", "M3H5K" ]
A 500 response body will be returned if any issues occur when calling the Lambda for example
- numberofrefs exceeds the limit
- numberofrefs parameter isn't an int
- any dynamodb Exceptions (key not found, unable to update, etc)
The number of references that can be returned in a single call is limited, because:
- prevent a single call using up all possible references;
- limits to the permitted size of the response.
Calling clients will need to handle this limit by making multiple calls to retrieve the required number of references if it is greater than the limit.
The limit is stored here for use by calling clients: reference_generator_limit
The DynamoDb stores the current counter used for generating unique references. The DynamoDb is encrypted so that it cannot be directly modified via the AWS console or AWS CLI. Below is an example of what the table looks like:
v1 | referenceCounter |
---|---|
fileCounter | 6 |
The DynamoDb is monitored by a specific set of Cloud Custodian rules:
Reference-counter-table-kms-key-check
: Checks if reference counter table encrypted with specific KMS key.Reference-counter-table-pitr-check
: Checks that point in time recovery (PITR) is enabled for reference counter table
Full details and code are here: Reference counter custodian rules
The API Gateway provides a RESTful interface to the Lambda function.
The urls for calling the reference generator can be obtained from the da-terraform-configurations
It can be called directly by providing the parameter numberofrefs={value}
by making a http request to one of the reference generator urls, for example:
https://j8ezi9m4z0.execute-api.eu-west-2.amazonaws.com/intg/counter?numberofrefs=2
.
The api gateway has a resource policy which restricts which services can call it.
Any new calling clients will need to provide an AWS IAM role which can call the API Gateway and this needs to be added to the API Gateway resource policy.
The references generated are based on the CTDb52 algorithm which is hosted on OCI tools scala. Every reference generated will have 'Z' prefixed to it, for example ["ZD4G25","ZD4G26"].
The last permitted reference that should be used is: "WWWWWWW" which corresponds to counter value 6103515624.
Deployment process of the service will depend on the hosting environment.
GitHub Actions workflow files should be prefixed with the hosting environment, eg tdr_
The service is currently hosted on Transfer Digital Records (TDR)
These are set in the tdr-terraform-github repository
MANAGEMENT_ACCOUNT
: AWS account number for the TDR management (mgmt) account.WORKFLOW_PAT
: GitHub access token for TDRSLACK_WEBHOOK
: TDR slack webhook
The reference counter in the DynamoDb table needs to be manually seeded before the table is encrypted.
- Create the DyanmoDb table using the Terraform but not adding the encryption
- Manually seed the counter in the table either via the AWS Cli or AWS Console:
{ "v1": { "S": "fileCounter" }, "referenceCounter": { "N": "0" } }
- Add table encryption using the Terraform
The Terraform directory contains the Terraform code used to create the DynamoDb, Lambda and API Gateway resources.
It relies on the da-terraform-configurations
and da-terraform-modules
projects.
-
Clone DA Terraform Configurations repository
[location of project] $ git clone https://github.com/nationalarchives/da-terraform-configurations.git
-
Clone DA Terraform Modules repository
[location of project] $ git clone https://github.com/nationalarchives/da-terraform-modules.git
Commit and push all the changes made in the terraform directory to its GitHub repo, then (in the GitHub repo):
Go the Actions tab -> Click ["Apply Terraform and deploy lambda"] -> Click "Run workflow" -> select the branch with the workflow file you want to use -> type the version to deploy -> Click the green "Run worklfow" button
Should the reference generator service need to be moved to a different hosting project then the following steps will need to be taken:
NOTE Before the move the current counter value will need to be noted down to ensure the new DynamoDb table is seeded with the correct counter to prevent duplicate references
- Add relevant Github Actions workflows for the new project to allow testing and deployment
- Update the da-terraform-configurations repo with the new values for the reference generator service in the relevant project file:
reference_generator_limit
reference_generator_intg_url
reference_generator_staging_url
reference_generator_prod_url
- Ensure the Cloud Custodian rules are implemented for on the new hosting project: DynamoDb > Security
- Set up relevant GitHub actions in the new hosting project: Deployment