Update Registry Settings
PUT/api/v33.02/settings/registry
x-prisma-cloud-target-env: {"permission":"policyContainers"}
Updates the registries to scan. The list of registries to scan is updated in a single shot.
To invoke this endpoint in the Console UI:
- Navigate to Defend > Vulnerabilities > Images > Registry settings.
- Under the Registries table, add a registry item using + Add registry
- Click the Save button.
Note: From 22.11 (Lagrange) release or later, you can add a maximum of 19,999 registry entries in Defend > Vulnerabilities > Images > Registry settings.
The API response returns an HTTP 400 error, if the number of registry specifications exceeds the maximum allowable limit of 19,999 registry entries.
General Set up and Scan Process
This endpoint works hand-in-hand with the /policies
endpoints.
To set up a registry for scanning:
-
Add your registry account information using this endpoint.
For example, specify the location and credentials of an ECR registry in your AWS account.
-
Prisma Cloud auto-discovers the images in the registries specified with this endpoint.
-
The list of auto-discovered images is passed to the scanner for evaluation.
The scanner uses the corresponding
/policies/vulnerability/images
and/policies/compliance/images
endpoints to assess each image.
cURL Request
Each registry to scan is specified as an item in the specifications
array.
Note: Submitting a PUT request with the specifications
array will first erase all the existing Registry entries, and then insert the new specifications
array from the PUT request.
An empty body will also erase all the existing Registry entries. For more information, see Remove a Registry.
We recommend that you send a GET scan settings request via Get Registry Settings and save the JSON response before sending a PUT API request to update the Registry Settings.
The critical fields for this endpoint are:
registry
- String specifying the registry URL.credentialID
- String specifying the registry credential.version
- String specifying the type of registry to scan and may be one of the following strings:
Version | Description |
---|---|
aws | Amazon EC2 Container Registry |
azure | Azure Container Registry |
2 | Docker Registry v2 |
dtr | Docker Trusted Registry |
gcr | Google Container Registry |
jfrog | JFrog Artifactory |
sonatype | Sonatype Nexus |
coreos | CoreOS Quay |
redhat | Red Hat OpenShift |
bluemix | IBM Cloud Container Registry |
The remaining fields in the specifications
object (e.g., repository
, exclusions
, etc.) are optional.
They let you refine the scope of what Prisma Cloud auto-discovers.
Note: An empty string in registry
implicitly refers to Docker Hub.
In repository
, use the library/
namespace to specify a Docker official image.
To see the current list of Docker official images, see here.
Set up a Private Registry for Scanning
Most registries you'll configure for scanning will be private. Prisma Cloud needs credentials to access private registries. To set this up:
- Create the credentials with the
/credentials
endpoint. - Retrieve the credential ID from the
/credentials
endpoint (_id
). - Create the registry setting with the recommended minimum required fields (
version
,registry
, andcredentialID
).
Example cURL Request
The following cURL command overwrites the current list of registries to scan with two new registries:
- The official Ubuntu 18.04 image in Docker Hub
- All repositories in a private AWS ECR registry
$ curl 'https://<CONSOLE>/api/v<VERSION>/settings/registry' \
-k \
-X PUT \
-u <USER> \
-H 'Content-Type: application/json' \
-d \
'{
"specifications": [
{
"version": "2",
"registry": "",
"repository": "library/ubuntu",
"tag": "18.04",
"os": "linux",
"cap": 5,
"credentialID": "<CREDENTIAL_ID1>",
"scanners": 2,
"collections": ["All"]
},
{
"version": "aws",
"registry": "<ACCOUNT_ID>.dkr.ecr.<REGION>.amazonaws.com",
"os": "linux",
"credentialID": "<CREDENTIAL_ID2>",
"scanners": 2,
"cap": 5,
"collections": ["All"]
}
]
}'
Note: No response will be returned upon successful execution.
Remove a Registry
To remove a registry from the list:
- Retrieve the current list using the GET method.
- Remove the entry from the
specifications
JSON array in the response. - Use the PUT method to submit the updated JSON object.
To delete all entries, submit an empty specifications
array. For example:
curl -k \
-u <USER> \
-H 'Content-Type: application/json' \
-X PUT \
-d '{"specifications":[]}' \
https://<CONSOLE>/api/v<VERSION>/settings/registry
Request
Responses
- 200
- default
OK