SWAG is a collection of repositories used to keep track of the metadata describing cloud accounts. Originally built to store data on AWS, swag is flexible enough to be used for multiple providers (AWS, GCP, etc.,).
For applications that manage or deploy to several different accounts/environments SWAG provides a centralized metadata store.
Managing your cloud accounts with SWAG allows for consuming application to focus on business logic, removing hardcoded lists or properties from your application.
These projects are part of the SWAG family and either manipulate or utilize SWAG data.
swag-api - Rest API and UI for SWAG Data
swag-functions - Transformation functions for SWAG Data
swag-client is available on pypi:
pip install swag_clientIf no options are passed to the SWAGManager it is assumed that you are using the file backend and the account namespace.
from swag_client.backend import SWAGManager
from swag_client.util import parse_swag_config_options
swag_opts = {}
swag = SWAGManager(**parse_swag_config_options(swag_opts))
account = swag.get_by_name('account')Configure SWAG by passing the client additional keyword arguments:
from swag_client.backend import SWAGManager
from swag_client.util import parse_swag_config_options
swag_opts = {
'swag.type': 's3',
'swag.bucket_name': s3_bucket_name
}
swag = SWAGManager(**parse_swag_config_options(swag_opts))Additionally SWAG can be used a singleton, allowing a single instance (and it's cache) to be shared across an application.
Declare swag in a file like extensions.py:
from swag_client.backend import SWAGManager
from swag_client.util import parse_swag_config_options
swag = SWAGManager()
When you're ready to configure swag:
from extensions import swag
swag_opts = {
'swag.type': 's3',
'swag.bucket_name': s3_bucket_name
}
swag.configure(**parse_swag_config_options(swag_opts))
Below is an example of the bare minimum JSON that will be created by SWAG with schema_version=2.
[{
"aliases": ["test"],
"contacts": ["admins@test.net"],
"description": "This is just a test.",
"email": "test@example.net",
"environment": "test",
"id": "012345678910",
"name": "testaccount",
"owner": "ExampleCorp",
"provider": "aws",
"sensitive": false
}]
Additionally SWAG has the ability to store metadata for services that may need to be tied to an account.
[{
"id": "012345678910",
...
"services": [
{
"name": "myService",
"metadata": {
"name": "testService"
},
"status": [
{
"region": "all",
"enabled": true
}
]
}
]
}]
This service metadata for serving as a single source of truth for applications that have to routinely manage/monitor/process multiple accounts.
Regardless of backend, SWAG uses the jmespath syntax for filtering data.
swag.get("[?id=='012345678910']") # valid jmespath filterMore information on jmespath filtering: http://jmespath.org/tutorial.html
SWAG also supports an older style calling convention. This convention only supports the S3 backend, additionally these functions are now deprecated and will be removed in the future.
from swag_client.swag import get_all_accounts
get_all_accounts(bucket='your-swag-bucket').get('accounts')or to filter by a service tag:
service = {'services': {'YOURSERVICE': {'enabled': True, 'randomflag': True}}}
get_all_accounts(bucket='your-swag-bucket', **service).get('accounts')All SWAG metadata is versioned. The most current version of the account metadata schema is v2. SWAG further provides the ability to transform data between schema version as necessary.
SWAG supports multiple backends, included in the package are file, s3, and dynamodb backends. Additional backends can easily be created and integrated into SWAG.
| Key | Type | Required | Description |
|---|---|---|---|
| swag.type | str | false | Type of backend to use (Default: 'file') |
| swag.namespace | str | false | Namespace for metadata (Default: 'accounts') |
| swag.schema_version | int | false | Schema version that will be returned to the caller. (Default: 'v2') |
| swag.cache_expires | int | false | Number of seconds to cache backend results (Default: 60) |
The S3 backend uses AWS S3 to SWAG metadata.
| Key | Type | Required | Description |
|---|---|---|---|
| swag.bucket_name | str | true | Raw S3 bucket name |
| swag.data_file | str | false | Full S3 key of file |
| swag.region | str | false | Region the bucket exists in. (Default: us-east-1) |
IAM Permissions required:
{
"Action": ["s3:GetObject"],
"Effect": ["Allow"],
"Resource: ["arn:aws:s3:::<swag-bucket-name>/<data-file>"]
}
If you wish to update SWAG metadata you will also need the following permissions:
{
"Action": ["s3:PutObject"],
"Effect": ["Allow"],
"Resource: ["arn:aws:s3:::<swag-bucket-name>/<data-file>"]
}
The file backend uses a file on the local filesystem. This backend is often useful for testing purposes but it not scalable to multiple clients.
| Key | Type | Required | Description |
|---|---|---|---|
| swag.data_dir | str | false | Directory to store data (Default: cwd()) |
| swag.data_file | str | false | Full path to data file |
The DynamoDB backend leverages AWSs DynamoDB as a key value store for SWAG metadata.
SWAG expects a Dynamodb Table already exists.
Create a new table via the AWS console named accounts with a primary key of id.
| Key | Type | Required | Description |
|---|---|---|---|
| swag.region | str | false | Region dynamodb table exists |
Note the above options except region is only needed if not SWAG table has been created.
Minimum Permissions required:
{
"Version": "2012-10-17",
"Statement": [
{
"Sid": "DescribeQueryTable",
"Effect": "Allow",
"Action": [
"dynamodb:DescribeTable",
"dynamodb:Query",
"dynamodb:Scan"
],
"Resource": "*"
}
]
}
If you wish SWAG to modify your table you will need the following additional permissions:
{
"Version": "2012-10-17",
"Statement": [
{
"Sid": "PutUpdateDeleteTable",
"Effect": "Allow",
"Action": [
"dynamodb:PutItem",
"dynamodb:UpdateItem",
"dynamodb:DeleteItem"
],
"Resource": "*"
}
]
}
Upon installation the swag_client creates a swag entrypoint which invokes the swag_client cli, example usage:
Examples:
swag --helpThe following describes the usage of all native fields included within the SWAG schema.
| Name | Type | Description |
|---|---|---|
| schemaVersion | int | Describes the current schema version |
| id | str | Unique ID of the account |
| name | str | Canonical name, according to the account naming standard |
| contacts | list(str) | List of team DLs that are majority stakeholders for the account |
| provider | str | One of: AWS, GCP, Azure |
| type | str | See schema context for field validation |
| status | list(dict) | See status schema |
| services | list(dict) | See service schema |
| environment | str | See schema context for field validation |
| sensitive | bool | Signifies if the account holds a special significance; (in scope for PCI, holds PII, contains sensitive key material, etc.,) |
| description | str | Brief description about the account's intended use. |
| owner | str | See schema context for field validation |
| aliases | list(str) | List of other names this account may be referred to as |
The V2 schema performs validation checks on certain fields to ensure values are within a defined list. Some of these are optional and configurable to allow users to specify values that make sense for their use case.
The allowed values for owner, environment and type can be set during SWAGManager initialization by passing a swag.schema_context object as part of the swag_opts.
If you do not specify a schema_context entry for a field then any value is permitted.
swag_opts = {
'swag.schema_context': {
'owner': ['netflix', 'dvd', 'aws', 'third-party'],
'environment': ['test', 'prod'],
'type': ['billing', 'security', 'shared-service', 'service']
}
}
| Name | Type | Description |
|---|---|---|
| name | str | Name of the service |
| regions | list(str) | List of regions - empty list indicates all regions |
| roles | list(dict) | List of roles that control access to this service. See role schema |
| metadata | dict | Service Level metadata |
| Name | Type | Description |
|---|---|---|
| region | str | Status per-region |
| status | str | One of: Created, In-progress, Ready, Deprecated, In-active, Deleted |
| notes | list(dict) | See notes schema |
| Name | Type | Description |
|---|---|---|
| date | date | Date note was created |
| text | str | Free text field with additional information |
| Name | Type | Description |
|---|---|---|
| id | str | Id of the role |
| policyUrl | str | URL with link to role permissions |
| roleName | str | Corresponding AWS role (if any) |
| googleGroup | str | Corresponding google group |
| secondaryApprover | str | DL needed to approve role |
Created - Account has been created but infrastructure has not yet been established
In-progress - Account infrastructure is currently being deployed
Ready - Account is ready for deployment
Deprecated - Account has been marked as deprecated, no new services should be deployed into this account
In-active - Account has been evacuated of all services
Deleted - Account has been marked as deleted
Billing Account(s) - The billing account in a multiple account architecture provides a central account for billing aggregation across the environment. Typically, no AWS resources (e.g. instances) run in the billing account.
Security Account(s) - The security account in a multiple account architecture provides environments for central log collection and analysis (e.g. CloudTrail, VPC Flow Logs) and security monitoring (e.g. configuration monitoring with Security Monkey, active security scanning and testing with other tools).
Shared Service Account(s) - Shared service accounts provide infrastructure, data, and services across the organization. All service/resource accounts will typically have network connectivity to shared service accounts.
Service/Resource Account Groups - Service/resource account groups host the bulk of systems and applications. Account groups are created based on various dimensions and will typically have production and test elements (separate accounts). Service/resource accounts may optionally have connectivity to other service/resource accounts.
See sample_accounts.json an example of the current json data created by SWAG.