Overview
Before beginning the installation process, ensure that the guidelines described in Installation Prerequisites have been fulfilled.
NOTE: Follow all procedures in order as they are presented here.
Installing the GitLab Integration
Configure a System Hook
Go to your GitLab instance and click Admin Area > System Hooks.
Fill in the fields according to these guidelines:
URL: The Mend webhook handler URL. The URL should end with '/payload'. Enter the webhook URL in the following format: http://<docker-wss-gls-app-destinationURL>:5678/payload.
Secret Token: Secret token for payload validation
Trigger: All checkboxes except Tag push events must be marked
Enable SSL Verification: Yes
Click Add system hook. Your system hook will appear in the System Hooks section at the bottom of the screen.
Creating a New GitLab User and a Personal Access Token
Create a new and regular GitLab user and name it @whitesource. This user does not need to be an admin.
NOTE: This user does not require a real email address. The Personal Access Token for the next step can be created by an admin impersonating the @whitesource user.Create a Personal Access Token for your new @whitesource user. Go to the @whitesource Settings > Access Tokens, and fill in the fields according to these guidelines:
Name: The name of your token, e.g. "MendToken"
Expires at (optional): The expiration date for your token
Scopes: All checkboxes should be marked.
Click Create personal access token. Copy the generated Personal Access Token to the clipboard, to be used in the next section.
Creating an Activation Key
Open a separate browser tab or window and log in to Mend.
Navigate to the Integrate page of the Mend application. Expand the Mend for GitLab Server bar to view the following fields:
GitLab Server API URL: Your GitLab Server instance's API URL. For example: https://GitLabDevServer.com/api/v4
GitLab Webhook URL: The URL of the Mend webhook handler (the same URL as the system hook from Configure a System Hook).
The webhook URL is used to create webhooks from GitLab projects the integration is installed for, to allow Mend Remediate to receive issue related events.
NOTE: If this webhook URL is on a local server, make sure your GitLab server is configured to allow outbound requests to local servers in Admin Area > Settings > Network > Outbound Requests. Here you can allow outbound requests to your entire local network or simply whitelist the Mend webhook URL.GitLab Webhook Secret: The webhook secret you entered when creating the system hook in Configure a System Hook.
GitLab Personal Access Token: The @whitesource user's personal access token created in the previous step.
Click Get Activation Key to generate your activation key. A new Service user is created for this integration inside the Mend Application with a WS prefix. NOTE: Do not remove this Service user and ensure this user remains part of the Admin group.
Copy the newly generated Activation Key (You will need it in the next procedure).
If you want Mend Remediate to only open automatic fix MRs in specific scenarios, clicking Manage Workflow Rules will allow you to configure custom rules for when Remediate opens MRs.
Running the UI Configuration Tool
The UI Configuration tool enables you to configure the deployment file according to your specific configuration requirements.
Open index.html located inside the wss-configuration directory via a Chrome or Firefox Web browser. The Mend Configuration Editor page is displayed.
Load the template JSON configuration file by clicking Choose File and selecting the file located at wss-configuration/config/prop.json. The General tab appears in the Editor.
Click the General tab and enter the Activation Key which you copied in the previous section.
To manage the creation of GitLab Issues or Commit Statuses globally across all of your organization's repositories, click the Issues tab. (The parameters in this tab are displayed in the table below.)
To display the Proxy tab, click the Advanced Properties checkbox on the Home tab. Proxy fields that are not mandatory (e.g., user name and password) must be left blank. (The parameters in this tab are displayed in the table below.)
Click Export, and save the JSON file with the name prop.json. This file will be used in the next sections.
You can export the JSON file at any time, even if you did not finish editing it in order to save your configurations and to enable assigning the configuration of a specific section to the appropriate professional in your organization (e.g., data source section may be assigned to the DBA of your organization).
When exporting the configuration file, it is important to give it the filename “prop.json”
Details on Attributes of the Configuration File
Section | Label | Name | Type | Mandatory | Description | Sample Value |
---|---|---|---|---|---|---|
General | Activation Key | bolt.op.activation.key | String | yes | Your generated activation key in the Mend application | |
Proxy | HTTP Proxy Host | proxy.host | Host Address | no | HTTP proxy host. Leave blank to disable. Default value: Empty | |
Proxy | HTTP Proxy Port | proxy.port | Integer | no | HTTP proxy port. Leave blank to disable. Default value: Empty | |
Proxy | Proxy User | proxy.user | String | no | Proxy Username (if applicable) | user |
Proxy | Proxy Password | proxy.password | String | no | Proxy Password (if applicable) | abc123 |
Advanced | Controller URL | controller.url | String | no | The ability to modify the App container URL in case its default name (wss-ghe-app) was modified. Default value: http://wss-ghe-app:5678 | |
Issues | Should Create Issues | bolt4scm.create.issues | Boolean | no | The ability to globally enable/disable Issues creation across all of your organization's repositories. Default value: true (NOTE: Supported from version 20.5.1.3 only) | |
Issues | Should Update Commit Status | bolt4scm.create.check.runs | Boolean | no | The ability to globally enable/disable commit statuses across all of your organization's repositories. Default value: true (NOTE: Supported from version 20.5.1.3 only) |
Uploading Mend Scan Results to the GitLab Security Dashboard (For GitLab Ultimate Users Only)
In order to upload the Mend scan results to the GitLab security dashboard, add the following YAML to your .gitlab-ci.yml file:
whitesource-security-publisher: image: openjdk:8-jdk when: manual script: - curl "{{WEBHOOK_URL}}/securityReport?repoId=$CI_PROJECT_ID&repoName=$CI_PROJECT_NAME&ownerName=$CI_PROJECT_NAMESPACE&branchName=$CI_COMMIT_REF_NAME&defaultBranchName=$CI_DEFAULT_BRANCH&commitId=$CI_COMMIT_SHA" -o gl-dependency-scanning-report-ws.json artifacts: paths: - gl-dependency-scanning-report-ws.json reports: dependency_scanning: - gl-dependency-scanning-report-ws.json expire_in: 30 days
Replace
{{WEBHOOK_URL}}
with the GitLab Webhook URL from Creating an Activation Key (remove the “payload” part from the URL).
NOTE: The CI job’s name must be whitesource-security-publisher in order for Mend to recognize and trigger it.
Running the Containers
Deploying Using Docker
Create a network bridge (this will create a private network between the different containers since all containers need to run within the same network):
docker network create -d bridge my_bridge
Run the wss-gls-app app container:
docker run --name wss-gls-app --network my_bridge -p 9494:9494 -p 5678:5678 -v <path/to/config/directory>:/etc/usr/local/whitesource/conf wss-gls-app:<version> # For example: docker run --name wss-gls-app --network my_bridge -p 9494:9494 -p 5678:5678 -v c:/tmp/gls/:/etc/usr/local/whitesource/conf/ wss-gls-app:19.9.1.1
Run the wss-scanner scanner container:
docker run --name wss-scanner-gls --restart=always --network my_bridge -p 9393:9393 -v <path/to/config/directory>:/etc/usr/local/whitesource/conf/ wss-scanner:<version> # For example: docker run --name wss-scanner-gls --restart=always --network my_bridge -p 9393:9393 -v c:/tmp/gls/:/etc/usr/local/whitesource/conf/ wss-scanner:19.9.1.1
Run the wss-remediate server container:
docker run --name remediate-server --restart=always --network my_bridge -e LOG_LEVEL=debug -p 8080:8080 -v <path/to/config/directory>/prop.json:/etc/usr/local/whitesource/conf/prop.json -v /tmp:/tmp wss-remediate:<version> # For example: docker run --name remediate-server --restart=always --network my_bridge -e LOG_LEVEL=debug -p 8080:8080 -v c:/tmp/gls/prop.json:/etc/usr/local/whitesource/conf/prop.json -v /tmp:/tmp wss-remediate:19.8.1 # NOTE: If port 8080 isn't available, you can use a different port by modifying the first port entry in the 'docker run' command. For example: # docker run --name remediate-server --network my_bridge -e LOG_LEVEL=debug -p 8082:8080 -v c:/tmp/gls/prop.json:/etc/usr/local/whitesource/conf/prop.json -v /tmp:/tmp wss-remediate:19.8.1
Deploying Using Helm Charts
The wss-deployment folder consists of the following structure:
helm
configs
templates
config.yaml
wssScmIntegration.yaml
Chart.yaml
values.yaml
Chart.yaml
This file contains information about the chart.
NOTE: Do not edit this file.
Values.yaml
This file represents the Mend integration image names and versions.
wsscanner: image: {image} version: {version} wsscontroller: image: {image} version: {version} wssremediate: image: {image} version: {version}
For each image declaration (wssscanner, wsscontroller, wssremediate), replace {image} and {version} with the actual built image name and version. NOTE: For wsscontroller, use the name and version of the wss-gls-app image.
An optional parameter, imagePullSecrets, can be added to this file in case Docker repository authentication is required.
configs/prop.json
Copy the helm folder from wss-deployment to your target environment. Inside the helm/configs folder, add the configuration properties JSON file (prop.json) that you previously edited and exported using the Configuration Editor.
templates/config.yaml
This is a configuration file pointing to the configs/prop.json file.
NOTE: Do not edit this file.
templates/wssScmIntegration.yaml
This is a configuration file containing all the parameters for deploying the integration.
NOTE: In this file, there are 3 dashes ("- - - ") that separate the services Do not remove them.
In order for the webhook URL to be accessible publicly by the integration, a load balancer service must be added to the file. An example of such a service is provided below:
apiVersion: v1 kind: Service metadata: name: lb1 namespace: acme annotations: external-dns.alpha.kubernetes.io/hostname: helm.acme.io service.beta.kubernetes.io/aws-load-balancer-backend-protocol: http service.beta.kubernetes.io/aws-load-balancer-ssl-ports: "443" service.beta.kubernetes.io/aws-load-balancer-ssl-negotiation-policy: "ELBSecurityPolicy-TLS-1-2-2017-01" service.beta.kubernetes.io/aws-load-balancer-ssl-cert: arn:aws:acm:us-east-7:834027593108:certificate/4720e07a-a231-4fd5-9c4a-12ab1450567d spec: type: LoadBalancer ports: - port: 443 name: https targetPort: 5678 selector: app: wss-controller
Selecting Repositories to Scan
If you are integrating multiple repositories and want to apply global configurations, refer here before continuing in this procedure.
Add the @whitesource user (the user you created during Creating a New GitLab User and a Personal Access Token) as a member with Maintainer permissions to the repositories you want Mend to scan. This is how Mend will determine which repositories will be scanned.
If you would like Mend for GitLab to scan an entire GitLab group, add the @whitesource user to the group to enable Mend for GitLab for all of the projects within that group.
NOTE: Adding the @whitesource user to a repository with any permissions less than Maintainer may create side-effects in the integration's functionality.Unless specified otherwise via the global configuration, an onboarding merge request is created for each repository to which the @whitesource user was added. This request contains a Mend configuration file (.whitesource) that can be customized before merging the request. The initial PR must be merged to the base branch first. This will then initiate the installation and start the first scan. You can then define further settings (like selected branches) in the .whitesource file.
In order to disable scanning for a repository, remove the @whitesource user from the repository members.
Accessing Scan Statistics via API
See here for more information.
Health Check APIs
See here for more information.
The .whitesource File
A Mend configuration file (.whitesource) is a JSON file added to each repository enabled for scanning. It provides configurable parameters for the Mend scan. The .whitesource file is only added in the default branch of the repository (unless modified, it is the master branch).
{ "scanSettings": { "configMode": "AUTO", "configExternalURL": "", "projectToken" : "", "baseBranches": [] }, "commitStatusSettings": { "displayMode": "diff", "vulnerableCommitStatus": "FAILED" }, "issueSettings": { "minSeverityLevel": "LOW", "openConfidentialIssues": true } } |
Parameters
Global Settings
Parameter | Type | Description | Required | Default |
---|---|---|---|---|
settingsInheritedFrom | String | When the global configuration is enabled, this parameter will specify the location of the whitesource-config repository from which it will inherit its configuration. It must contain the GitLab user name, repository name and branch (optional) of the repo-config.json file location. The default branch is 'master', but can be modified according to the location of the repo-config.json file in the whitesource-config repo. NOTE: You can override specific parameters that are relevant only in the specific repository by adding these after this parameter. Parameters with type of array do not override the value from global configuration, but only add new values. Examples: Using only values defined in the global configuration: "settingsInheritedFrom": "whitesource-config/whitesource-config@master" Using values defined in the global configuration and overriding the scan settings parameters: "settingsInheritedFrom": "whitesource-config/whitesource-config@master", "scanSettings": { "projectToken": "12345", "baseBranches": ["master","integration"] } | No | N/A |
overrideConfigAllowList | Array | When the global configuration is enabled, this parameter will regulate the ability of repositories that inherit their configuration from the whitesource-config repository to override the parameters locally. There are three options:
NOTE: This parameter must be used in the repo-config.json file of the whitesource-config repository. | No | null |
Scan Settings (scanSettings)
Parameter | Type | Description | Required | Default |
---|---|---|---|---|
configMode | String | The configuration mode to be used for each scan. There are three options:
| No | Auto |
configExternalURL | String | The URL of the external configuration file (you can choose any filename). The configuration file content should be in the same format as the Unified Agent configuration file. The following protocols are supported: 'ftp://', 'http://', 'https://'. For example: 'https://mydomain.com/whitesource-settings/wss-unified-agent.config' NOTE: This parameter is relevant only if configMode was set to EXTERNAL. | No | Empty |
projectToken | String | Adds the ability to map a GitLab repository to a Mend project. The parameter used needs to be the Mend project token. NOTE: Not supported in the Global Configuration. | No | Empty |
baseBranches | Array | Adds the ability to specify one or more base branches for which scanning results will be sent to a new Mend project. Example usage: ["master", “integration"] This will set both master and integration branches as base branches. Note the following:
NOTE: This parameter is available only from version 20.7.1. | No | Empty In this case, the base branch only consists of the default branch. |
enableLicenseViolations | Boolean | When enabled, a new Mend License Check will be generated for each valid push. NOTES:
| No | false |
Commit Status Settings (commitStatusSettings)
Parameter | Type | Description | Required | Default |
---|---|---|---|---|
displayMode | String | How to display Mend security information for a scan performed on a non-base branch:
| No | diff |
vulnerableCommitStatus | String | Customizable commit status settings.
| No | FAILED |
licenseCommitStatus | String | Customizable commit status settings. NOTE: The license check is dependent on the vulnerabilities check and will not be triggered if vulnerableCommitStatus is set to none.
| No | FAILED |
iacCommitStatus | String | Customizable commit status settings.
| No | FAILED |
showWsInfo | Boolean | Whether to show additional Mend information such as the project token inside the Mend Commit Status (after the scan token). Mend information is only displayed if the commit originated from a base branch. The following hidden JSON object will also be added inside the Commit Status when this parameter is enabled: <!-- <INFO>{"projectToken":"1cd2d2a8651145c087609e0a43f783e95f7008cb908541498348fed529572e01"}</INFO> --> NOTE: Additional Mend data may be added inside the JSON object in the future. | No | false |
Issue Settings (issueSettings)
Parameter | Type | Description | Required | Default |
---|---|---|---|---|
minSeverityLevel | String | Enables users to decide whether to open a new Issue only if a certain severity level is available on a detected vulnerability. Available values for minSeverityLevel:
NOTE: The Mend Security Check summary is also affected by this parameter. | No | LOW |
openConfidentialIssues | Boolean | Whether the GitLab issues opened by Mend will be confidential issues. | No | false |
displayLicenseViolations | Boolean | Whether to generate an Issue for every detected license policy violation. NOTE: This parameter is relevant only if enableLicenseViolations (scanSettings) is set to true. | No | true (only if enableLicenseViolations (scanSettings) is set to true) |
issueType | String | Defines the type of issue that will be created in the repository: VULNERABILITY (one for each vulnerability) or DEPENDENCY (one for each direct dependency). | No | DEPENDENCY |
customLabels | Array | Setting that will define labels that will be added to the issues created after the scan. Usage example: { "issueSettings": { "customLabels": ["label1","label2"] } } Following labels are not available for the use:
| No | Empty |
Remediate Settings (remediateSettings)
Parameter | Type | Description | Required | Default |
---|---|---|---|---|
enableRenovate | Boolean | When enabled, Remediate will raise automated Merge Requests for outdated dependencies in addition to Merge Requests remediating vulnerable dependencies. Remediate will then perform all the functionality and support all the configuration options available in Mend Renovate. See Renovate configuration options for all configuration options. Refer here for parameter usage. | No | false |
transitiveRemediation | Boolean | Whether to enable transitive remediation for NPM repos. When npm v6 (npm v7 is not currently supported) is used with a package-lock.json file, and vulnerabilities are found within transitive dependencies in the file, then in most cases Remediate is able to successfully remediate the vulnerability. Sometimes it may not be possible to successfully remediate because a parent dependency does not yet have a new release that allows the necessary fixed-in version of the transitive dependency. | No | false |
workflowRules | Object | This parameter is used to specify the rules that regulate when to open remediation pull requests. Usage examples: "remediateSettings": { "workflowRules": { "enabled": true, "minVulnerabilitySeverity": "LOW" } } "remediateSettings": { "workflowRules": { "enabled": true, "minVulnerabilityScore": 1.5, "maxVulnerabilityScore": 10 } } | Yes | "workflowRules": { "enabled": true } |
workflowRules.enabled | Boolean | Enables Workflow Rules being set from a .whitesource file. Note: workflow rules can also be set in the Mend application in the Admin → Integration Workflow Rules. But if this parameter is set to | Yes | true |
workflowRules.minVulnerabilitySeverity | String | The minimal vulnerability severity level to automatically create remediation pull requests for. Allowed values - E.g. if set to Note: if this parameter is used together with minVulnerabilityScore and maxVulnerabilityScore than only minVulnerabilitySeverity will have affect. | No | LOW |
workflowRules.minVulnerabilityScore | Float | The minimal vulnerability CVSS 3 score to automatically create remediation pull requests for. Allowed values - floats with one decimal from 0 to 10. For more information on CVSS 3 Scores, click here. Note: if this parameter is used together with minVulnerabilitySeverity it will not have any effect. | No | 0 |
workflowRules.maxVulnerabilityScore | Float | The maximal vulnerability CVSS 3 score to automatically create remediation pull requests for. Allowed values - floats with one decimal from 0 to 10. For more information on CVSS 3 Scores, click here. Note: if this parameter is used together with minVulnerabilitySeverity it will not have any effect. | No | 10 |
Private Registry Settings (hostRules)
Parameter | Type | Description | Required | Default |
---|---|---|---|---|
matchHost | String | Defines where the credentials will be applied during the scan. If you want to apply credentials only for a nested path within a host, then write If the same credentials apply to all paths on a host and not on any subdomains, configure Finally, to apply credentials to all hosts within the domain, use a | No | Empty |
hostType | String | Type of private registry. Supported values: Required if | No | Empty |
username | String | Used when credentials consist of username and password. | No | Empty |
password | String | Used when credentials consist of username and password, should be encrypted by this instruction. Encrypted secret that will be applied as a credential to the host set in the "encrypted": { "password": "3f832f2983yf89hsd98ahadsjfasdfjaslf............" } | No | Empty |
token | String | Used when credentials consist of username and password, should be encrypted by this instruction. Encrypted secret that will be applied as a credential to the host set in the "encrypted": { "token": "3f832f2983yf89hsd98ahadsjfasdfjaslf............" } | No | Empty |
Providing a Global Configuration File
NOTE: Supported from version 20.5.1.3 only.
You can provide a custom .whitesource configuration file as part of the wss-gls-app container, in order to apply it globally to all of your organization's repositories. Doing so will apply the file to all onboarding pull requests for newly-selected repos. Repos which were already selected and activated before this change will not be affected by this global configuration. Only newly onboarded repos will be affected.
To apply this global change, do as follows:
Stop the wss-gls-app container.
In the "wss-gls-app/conf" folder, add your custom “.whitesource” file (where the prop.json file is located).
Start the wss-gls-app container.
Configuration Error Issues
Will alert the user on configuration errors that affects their scan by creating a configuration error issue and commit status. In case of such an error the following will occur:
Stop the workflow. Do not create a scan or the Mend Security commit status.
Create a “Configuration Failed” commit status.
For each config file that failed parsing - create a new type of issue, titled Action Required: Fix Mend Configuration File - {fileName}. If the error originated from the repo-config.json or global-config.json files, then the issue will be created in the whitesource-config repo.
Handled errors:
Error parsing the configuration files (.whitesource/repo-config.json/global-config.json json)
Missing repository and/or branch in the inheritance configuration
Handling Private Registries and Authenticated Repositories
Private registries hosted on any platform that can be accessed with credentials are supported (Nexus, Artifactory, JFrog, etc.)
Supported languages and package managers:
NPM
Yarn
Maven
Gradle
PIP
Go
Nuget
Ruby
In order to scan dependencies from private registries and authenticated repositories, Mend must be provided with credentials, such as an NPM token. These credentials must be added as encrypted secrets to the .whitesource file, either per-repository or in the shared global config, if the secret scope is org-wide.
Сreate the encrypted secrets. Each secret you encrypt must be scoped to a GitLab group or repository and use of it will be restricted to those within the app.
Use GPG to generate a PGP Key. Use the command
gpg --full-generate-key
and follow the prompts to generate a key. Please note that at this time we do not support using a passphrase for decryption, so it is best to generate the keys without a passphrase. Name and email are not important.Copy the key ID from the output or run
gpg --list-secret-keys
if you forgot to take a copy. This is your public key.Run
gpg --armor --export-secret-keys YOUR_NEW_KEY_ID > ws-private-key.asc
to generate an armored (text-based) private key fileRun
gpg --armor --export YOUR_NEW_KEY_ID > ws-public-key.asc
to generate an armored (text-based) public key file
Provide the private key to the Controller, Remediate, and Scanner with environmental variable (learn more about environmental variables in the Advanced Technical Information documentation). There are two options for how to do it, but only one option should be used.
WS_HOST_RULES_PRIVATE_KEY
- the value of the private key itself.WS_HOST_RULES_PRIVATE_KEY_FILE_PATH
- path to the file containing the private key. This file should be mapped to the running containers.
Open index-enterprise.html in your favorite editor.
Find and replace the text "COPY_YOUR_PUBLIC_PGP_KEY_HERE" with your newly generated public key and save the file.
const publicKeyString = `COPY_YOUR_PUBLIC_PGP_KEY_HERE`;
Generate a secret. There are the following fields on the encryption page:
Organization\Group - required, your Gitlab group to which tokens secret be scoped.
Repository - optional, your Gitlab repository to which secret should be scoped.
Raw value - required, confidential values/secrets such as tokens or passwords.
Encrypted value - the result of the encryption to be used in the integration.
After the secret is created, please add it to the hostRules parameter of the .whitesource file.
Example of hostRules:
{ "hostRules": [ { "matchHost": "registry.npmjs.org", "hostType": "npm", "encrypted": { "token": "3f832f2983yf89hsd98ahadsjfasdfjaslf............" } }, { "matchHost": "https://custom.registry.company.com/maven/", "hostType": "maven", "username": "bot1", "encrypted": { "password": "p278djfdsi9832jnfdshufwji2r389fdskj........." } } ] }
NOTE:
Copy the entire output of the key generator including comments to paste into the string.
i.e. include "-----BEGIN PGP PUBLIC KEY BLOCK-----..."
The string uses javascript backticks and not quotes. This is to allow a multi-line string so that you do not have to replace any line breaks with new-line characters. Be aware of any auto indenting by your editor that may introduce spaces to the public key and cause encryption to fail.
We use asymmetric public-key cryptography of the PGP methodology. Organization/Group, Repository, Raw Value - all information you provide on the encryption page is secured with this approach.