Docker Scout image analysis

Megha SharmaMegha Sharma
6 min read

When you activate image analysis for a repository, Docker Scout automatically analyzes new images that you push to that repository.

Image analysis extracts the Software Bill of Material (SBOM) and other image metadata, and evaluates it against vulnerability data from security advisories.

If you run image analysis as a one-off task using the CLI or Docker Desktop, Docker Scout won’t store any data about your image. If you enable Docker Scout for your container image repositories however, Docker Scout saves a metadata snapshot of your images after the analysis. As new vulnerability data becomes available, Docker Scout recalibrates the analysis using the metadata snapshot, which means your security status for images is updated in real-time. This dynamic evaluation means there’s no need to re-analyze images when new CVE information is disclosed.

Docker Scout image analysis is available by default for Docker Hub repositories. You can also integrate third-party registries and other services.

πŸ‘‰ Activate Docker Scout on a repository

The free tier of Docker Scout lets you use Docker Scout for up to 3 repositories per Docker organization. You can update your Docker Scout plan if you need additional repositories.

Before you can activate image analysis on a repository in a third-party registry, the registry must be integrated with Docker Scout for your Docker organization. Docker Hub is integrated by default.

To activate image analysis:

  1. Go to Repository settings in the Docker Scout Dashboard.

  2. Select the repositories that you want to enable.

  3. Select Enable image analysis.

If your repositories already contain images, Docker Scout pulls and analyzes the latest images automatically.

Note: You must have the Editor or Owner role in the Docker organization to activate image analysis on a repository.

πŸ‘‰ Analyze registry images

To trigger image analysis for an image in a registry, push the image to a registry that’s integrated with Docker Scout, to a repository where image analysis is activated.

Note: Image analysis on the Docker Scout platform has a maximum image file size limit of 10 GB, unless the image has an SBOM attestation.

🌟 Sign in with your Docker ID, either using the docker login command or the Sign in button in Docker Desktop.

🌟 Build and push the image that you want to analyze.

$ docker build --push --tag <org>/<image:tag> --provenance=true --sbom=true .

Building with the --provenance=true and --sbom=true flags attaches build attestations to the image. Docker Scout uses attestations to provide more fine-grained analysis results.

🌟 Go to the Images page in the Docker Scout Dashboard.

The image appears in the list shortly after you push it to the registry. It may take a few minutes for the analysis results to appear.

πŸ‘‰ Analyze images locally

You can analyze local images with Docker Scout using Docker Desktop or the docker scout commands for the Docker CLI.

Docker Desktop:

To analyze an image locally using the Docker Desktop GUI:

  1. Pull or build the image that you want to analyze.

  2. Go to the Images view in the Docker Dashboard.

  3. Select one of your local images in the list.

CLI:

The docker scout CLI commands provide a command line interface for using Docker Scout from your terminal.

  • docker scout quickview: summary of the specified image.

  • docker scout cves: local analysis of the specified image.

  • docker scout compare: analyzes and compares two images.

By default, the results are printed to standard output. You can also export results to a file in a structured format, such as Static Analysis Results Interchange Format (SARIF).

Quickview:

The docker scout quickview command provides an overview of the vulnerabilities found in a given image and its base image.

$ docker scout quickview traefik:latest
βœ“ SBOM of image already cached, 311 packages indexed

Your image  traefik:latest  β”‚    0C     2H     8M     1L
Base image  alpine:3        β”‚    0C     0H     0M     0L

If your the base image is out of date, the quickview command also shows how updating your base image would change the vulnerability exposure of your image.

$ docker scout quickview postgres:13.1
βœ“ Pulled
βœ“ Image stored for indexing
βœ“ Indexed 187 packages

Your image  postgres:13.1                 β”‚   17C    32H    35M    33L
Base image  debian:buster-slim            β”‚    9C    14H     9M    23L
Refreshed base image  debian:buster-slim  β”‚    0C     1H     6M    29L
                                          β”‚    -9    -13     -3     +6
Updated base image  debian:stable-slim    β”‚    0C     0H     0M    17L
                                          β”‚    -9    -14     -9     -6

CVEs:

The docker scout cves command gives you a complete view of all the vulnerabilities in the image. This command supports several flags that lets you specify more precisely which vulnerabilities you're interested in, for example, by severity or package type:

$ docker scout cves --format only-packages --only-vuln-packages \
  --only-severity critical postgres:13.1
 βœ“ SBOM of image already cached, 187 packages indexed
 βœ— Detected 10 vulnerable packages with a total of 17 vulnerabilities

 Name         Version                Type        Vulnerabilities
───────────────────────────────────────────────────────────────────────────
 dpkg         1.19.7                 deb       1C     0H     0M     0L
 glibc        2.28-10                deb       4C     0H     0M     0L
 gnutls28     3.6.7-4+deb10u6        deb       2C     0H     0M     0L
 libbsd       0.9.1-2                deb       1C     0H     0M     0L
 libksba      1.3.5-2                deb       2C     0H     0M     0L
 libtasn1-6   4.13-3                 deb       1C     0H     0M     0L
 lz4          1.8.3-1                deb       1C     0H     0M     0L
 openldap     2.4.47+dfsg-3+deb10u5  deb       1C     0H     0M     0L
 openssl      1.1.1d-0+deb10u4       deb       3C     0H     0M     0L
 zlib         1:1.2.11.dfsg-1        deb       1C     0H     0M     0L

πŸ‘‰ Vulnerability severity assessment

Docker Scout assigns a severity rating to vulnerabilities based on vulnerability data from advisory sources. Advisories are ranked and prioritized depending on the type of package that’s affected by a vulnerability.

For example, if a vulnerability affects an OS package, the severity level assigned by the distribution maintainer is prioritized.

If the preferred advisory source has assigned a severity rating to a CVE, but not a CVSS score, Docker Scout falls back to displaying a CVSS score from another source. The severity rating from the preferred advisory and the CVSS score from the fallback advisory are displayed together. This means a vulnerability can have a severity rating of LOW with a CVSS score of 9.8, if the preferred advisory assigns a LOW rating but no CVSS score, and a fallback advisory assigns a CVSS score of 9.8.

Vulnerabilities that haven’t been assigned a CVSS score in any source are categorized as Unspecified (U).

Docker Scout doesn’t implement a proprietary vulnerability metrics system. All metrics are inherited from security advisories that Docker Scout integrates with. Advisories may use different thresholds for classifying vulnerabilities, but most of them adhere to the CVSS v3.0 specification, which maps CVSS scores to severity ratings according to the following table:

CVSS score         Severity rating
──────────────────────────────────
0.1 – 3.9            Low (L)
4.0 – 6.9            Medium (M)
7.0 – 8.9            High (H)
9.0 – 10.0           Critical (C)

Note: that, given the advisory prioritization and fallback mechanism described earlier, severity ratings displayed in Docker Scout may deviate from this rating system.

πŸ‘‰ Maximum image size:

Image analysis on the Docker Scout platform, and analysis triggered by background indexing in Docker Desktop, has an image file size limit of 10 GB (uncompressed). To analyze images larger than that, you can either:

  • Attach SBOM attestations at build-time

  • Use the CLI to analyze the image locally

Images analyzed locally with the CLI and images with SBOM attestations have no maximum file size.

0
Subscribe to my newsletter

Read articles from Megha Sharma directly inside your inbox. Subscribe to the newsletter, and don't miss out.

Written by

Megha Sharma
Megha Sharma

πŸ‘‹ Hi there! I'm a DevOps enthusiast with a deep passion for all things Cloud Native. I thrive on learning and exploring new technologies, always eager to expand my knowledge and skills. Let's connect, collaborate, and grow together as we navigate the ever-evolving tech landscape! SKILLS: πŸ”Ή Languages & Runtimes: Python, Shell Scripting, YAML πŸ”Ή Cloud Technologies: AWS, Microsoft Azure, GCP πŸ”Ή Infrastructure Tools: Docker, Terraform, AWS CloudFormation πŸ”Ή Other Tools: Linux, Git and GitHub, Jenkins, Docker, Kubernetes, Ansible, Prometheus, Grafana