Skip to content

Latest commit

 

History

History
135 lines (91 loc) · 6.19 KB

09-security_configuration-management.md

File metadata and controls

135 lines (91 loc) · 6.19 KB
Raw

Configuration management

The goal: introduction into configuration-management

Disclaimer: the following section give only a brief introduction into the topics and propose a few low-hanging fruits to get started.

Common Vulnerabilities and Exposures (CVE)

  1. (Almost) all the software has issues (intentional or not).

    • Likely, we are not eager to run application with known issues inside.
    • When an issue is discovered, engineers try to fix it and release a fix/patch.
    • When a patch is released, we have to apply it to get rid of the issue.

  2. Common Vulnerabilities and Exposures system provides a standard to describe vulnerable software: ID, description, severity, references.

  3. There are tools that helps to compare your software with CVE database to find if your software has any known issue. You can scan your OS, application codebase, container images

  4. One of popular scanners is trivy. You can install it to your machine or run inside a container, for example:

    docker run --rm -v ~/.trivy:/root/.cache/ aquasec/trivy:0.50.1 image python:3.12

    Discalimer: there are a lot of other free and commercial scanners: Clair, anchore/grype, snyk to name a few.

    NOTE: trivy does not actually dive into packages content, but trusts reported version number. For example, you can update java pom file manually to trick the tool. Thus it's not a reliable runtime protection, but an assistant to understand your software.

  5. You can include image-scanning as gate before pushing an image to Container Registry. thus, if container has a known vulnerability - image is not released:

    docker run --rm -v ~/.trivy:/root/.cache/ aquasec/trivy:0.50.1 image --exit-code 1 --no-progress python:3.12

# get exit-code of last command
echo $?

docker run --rm -v ~/.trivy:/root/.cache/ aquasec/trivy:0.50.1 image --exit-code 1 --no-progress alpine:latest

echo $?

    You can fail only on critical issues --severity HIGH,CRITICAL, ignore unfixed --ignore-unfixed or particular CVEs --ignorefile ~/.trivyignore.

  6. Often, you do not have to fix all the CVEs manually, but change your base image:

    ### buster
docker run --rm -v ~/.trivy:/root/.cache/ aquasec/trivy:0.50.1 image --exit-code 1 --no-progress python:3.12

python:3.12 (debian 12.5)
=========================
Total: 988 (UNKNOWN: 5, LOW: 525, MEDIUM: 350, HIGH: 103, CRITICAL: 5)
    ### buster-slim
docker run --rm -v ~/.trivy:/root/.cache/ aquasec/trivy:0.50.1 image --exit-code 1 --no-progress python:3.12-slim

python:3.12-slim (debian 12.5)
==============================
Total: 116 (UNKNOWN: 0, LOW: 71, MEDIUM: 33, HIGH: 11, CRITICAL: 1)
    ### alpine
docker run --rm -v ~/.trivy:/root/.cache/ aquasec/trivy:0.50.1 image --exit-code 1 --no-progress python:3.12-alpine

python:3.12-alpine (alpine 3.19.1)
==================================
Total: 0 (UNKNOWN: 0, LOW: 0, MEDIUM: 0, HIGH: 0, CRITICAL: 0)

    The difference in base images is also amount of used tools: more tools - more issues. For example, compare python image sizes:

    $ docker images python
REPOSITORY   TAG           IMAGE ID       CREATED      SIZE
python       3.12          099bf23b94d9   6 days ago   1.02GB
python       3.12-slim     0e42464fe231   6 days ago   130MB
python       3.12-alpine   f44387b48281   6 days ago   57.1MB

Software supply-chain protection

  1. Software supply chain (e.g. code delivery process): write code, build, test, deploy, run, monitor.

  2. Build your solution only on top of trusted sources. For example, do not use no-name pip/npm packages or base docker-images. Consider pulling used dependencies to your local package registry instead of pulling from public sources.

  3. During writing code you can use Static Code Analysis tools to catch issues earlier. To find a tool for your language, search static code analysis {your programming language}. Maybe, the most popular one is Sonarqube.

  4. Regularly patch your dependencies and run scanners, like trivy above.

  5. Sign your commits, container images (or with cosign).

  6. Verify signatures. For example:

  7. Protect access to your build tools.

  8. Use immutable versions whenever is possible.

    For example, container images has immutable id:

    # run image using mutable tag:
docker run -it --rm python:3.12

# find image immutable digest:
docker inspect --format='{{index .RepoDigests 0}}' python:3.12
# or
docker images --digests

# run image by digest instaed of tag:
# NOTE: likely, at the moment the hash is different, e.g. someone pushed a new image with the same 3.9 tag
docker run -it --rm python@sha256:e0e2713ebf0f7b114b8bf9fbcaba9a69ef80e996b9bb3fa5837e42c779dcdc0f

Summary

  1. Start from trusted code
  2. Scan and sign your code and container images
  3. Regularly patch your software
  4. Less CVEs - less issues
  5. Smaller image -> less tools -> less CVEs
  6. Use immutable versions

Further reading