The Anchore Plugin offers several operations to enhance container scanning in your devops pipeline. Anchore is specifically targeted at image scanning, unlike some of the other tools such as Docker Bench Security, which make it a perfect match to your image building CI process.
For those inexperienced with Anchore, the scanLocalImage operation is just where to start. This operation doesn't require any prerequisites other than having docker installed. For those looking scan and publish results to an existing Anchore Engine installation, analyzeLocalImage and analyzeImage are what you are looking for.
Anchore Inline Scanner
This plugin uses Anchore CI Tools to perform local scanning. Anchore requires images to be pushed to a registry prior to analyzing. The Anchore CI Tools get around this by starting a temporary local Anchore container with a 'localbuild' registry. This allows you to scan and analyze images directly after building on your local docker engine.
Supported Versions
Docker version 18.06+ (Older versions of Docker may work but have not been tested)
Windows is NOT supported
Key Features
Can be used without any previous Anchore installation or knowledge.
Supports both inline scanning (scanLocalImage) and publishing results to an existing installation (analyzeImage and analyzeLocalImage).
Ability to use predefined or custom policy bundles to define what should be scanned or analyzed.
Optional groovy script validation with variable support for the scan results as well as FlexDeploy environment variables.
Outputs and Groovy Condition
Each operation in this plugin supplies the option to configure a Groovy script to determine if the scan results are inacceptable and should fail the operation. Below is a list of variables available to the Groovy script.
Outputs
Each of these variables is also available as a plugin output
Groovy Variable | Plugin Output | Description | Example |
---|---|---|---|
STATUS |
| Overall status of the scan/analysis. This is ultimately determined by the policy bundle that is used. Possible values are pass or fail. | fail |
FINAL_ACTION |
| Similar to the status, this provides the recommended action for the image. | stop |
STOP_COUNT |
| An integer count of the number of STOP rules detected for the image | 2 |
WARN_COUNT |
| An integer count of the number of WARN rules detected for the image | 10 |
ALL_COUNTS |
| A full list of all counts returned | [STOP:2, WARN:10, GO:37] |
All other FlexDeploy Variables are also available in the Groovy Script Condition |
Example Groovy Script Conditions
The most common use case lets the policy bundle unequivocally determine failure.
STATUS == "fail"
No warn or stop checks found
(STOP_COUNT + WARN_COUNT) != 0
A script that is more lenient on non-master stream builds (could be only master stream is pushed to the registry and other streams are for local testing).
//master is not allowed any stop checks whereas every other stream can have 5 or less FDBLD_STREAM_NAME == "master" ? STOP_COUNT > 0 : STOP_COUNT > 5
Project Configuration
Anchore is FlexDeploy's preferred choice when it comes to container scanning. As such, you can configure your project to scan the specified image without any configuration other than setting the below properties:
For more information on the properties outlined above and all of the Project Container Configuration see here.