cnfuzz

module
v0.2.1 Latest Latest
Warning

This package is not in the latest version of its module.

Go to latest
Published: Dec 8, 2022 License: Apache-2.0

README

Go GitHub Action Container Build GitHub Action Kubernetes Integration GitHub Action

cnfuzz - Cloud Native Web API Fuzzer

"Breaking Cloud Native Web APIs in their natural habitat."

Fuzzing web APIs in their fully converged Cloud Native state renders more representative results, just like it would have been deployed in production.

cnfuzz is a project written in Golang that automates fuzzing web APIs deployed in Kubernetes clusters. By tracking hashes of all container images, (re)deployed web API versions will be fuzzed - to detect potential security and stability issues and stores its results in a data lake.

By using Pod Annotations, OpenAPI and RESTler by Microsoft both discovery and fuzzing is being completely automated.

Why?

  • You want to fuzz web API logic where they actually operate, especially when fuzzing complete Service Meshes
  • You want to integrate and/or build data lakes with fuzzing data on top of AWS S3 based storage
  • You want to save expensive Cloud CI/CD pipeline credits by using idle Kubernetes cluster resources
  • You want fuzzing te be done outside of your CI/CD pipeline
  • You have heavy performance requirements for your fuzzing and Cloud based CI/CD pipelines do not suffice
  • You want to fuzz web API's of services which are interconnected and are being deployed by different teams in the same Kubernetes cluster
  • You want to automatically fuzz existing opensource software for instability issues
  • You get excited over fuzzing farms

Architecture

Usage

Installation

helm repo add cnfuzz https://suecodelabs.github.io/cnfuzz
helm repo update
helm install cnfuzz cnfuzz/cnfuzz

Getting started

All it takes to getting started after installation of cnfuzz on your Kubernetes cluster is to annotate your Kubernetes Pods, Deployments, DaemonSets, etc like in the following example:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: my-api
spec:
  selector:
    matchLabels:
      app: my-api
  replicas: 1
  template:
    metadata:
      labels:
        app: my-api
      annotations:
        cnfuzz/enable: "true"
        cnfuzz/open-api-doc: "/swagger/swagger.json"
        cnfuzz/secret: "0d5989ed-d60c-470e-b1b5-576fcf0f5d8c"
    spec:
      containers:
        - name: myapi
          image: my-api
          imagePullPolicy: Always
          ports:
            - containerPort: 80

Development

Setup Kubernetes development environment

Build and run

Kind
# don't forget to commit your changes locally before deploying to Kind.
make kind-init
Rancher Desktop
# don't forget to commit your changes locally before deploying to Rancher Desktop.
make rancher-init

These commands do the following:

  • Setup initial deployment of cnfuzz
  • Build a container image and load it into Kubernetes
  • Install cnfuzz via helm with the local built image
  • Create example webapi deployment to fuzz
Kind
# don't forget to commit your changes locally before deploying to Kind.
make kind-build
Rancher Desktop
# don't forget to commit your changes locally before deploying to Kind.
make rancher-build

These commands do the following:

  • (re)-build cnfuzz and upgrade deployment with latest image
Cleanup the build

This command does the following:

  • delete the cnfuzz installation from Kubernetes
make k8s-clean
# If you did a git pull between the above build and the below k8s-clean
# you will see an error. Specify the release as follows:
make k8s-clean GIT_COMMIT=f4fd3d2

Build project

For building the project you can use the Makefile.

Build Docker image
IMAGE=myrepo/cnfuzz make image
Compile binary
# Compile project to binary dist/cnfuzz
make build

Debugging

Useful flags for debugging:

--debug # cnfuzz outputs extra logging
--inside-cluster=false # cnfuzz will use your local config in $HOME/.kube/config (by default)
--restler-time-budget 0.001 # RESTler jobs complete almost instantly
--cache in_memory # use in memory cache for fuzzed images, cache gets deleted when cnfuzz exits

The code can be debugged in your IDE (outside the cluster) with the --inside-cluster=false flag. But you can also attach a debugger to a running pod inside a cluster using DevSpace.

  1. Start by installing DevSpace
  2. Run devspace dev in the root directory of this repository
  3. Run air -c air.toml inside the container
  4. Edit the code and set breakpoints
  5. Attach your IDE to the debugger inside the container

Prepare for release

cd docs
helm package ../charts/cnfuzz
helm repo index --url https://suecodelabs.github.io/cnfuzz/ .

Roadmap

  • Opensource graduation research project ❤️
  • Support Kustomize
  • Integrate more tightly with Kubernetes
  • Autodiscovery of possible URI prefixes

Sponsors

Engineering Team

  • Luuk van den Maagdenberg, Lead Developer
  • Pim Merks, Developer
  • Robert Scholts, Developer
  • Sylvia van Os, Developer
  • Ofer Chen, Developer
  • Serge van Namen, Developer / Community Lead
  • Hans Strijker, Maintainer
  • Sam Crauwels, Maintainer
  • Michiel Westerink, Maintainer

Contribution

Create an issue, open up a PR or contact us via engineering@sue.nl

Community

Every first Thursday of the month at 18:30 CET there will be a hybrid community meetup for users, developers and maintainers of the project hosted at Sue B.V. in the Netherlands.

Google Meet or see you at Sue B.V. including lovely drinks and food before the meetup!

Swag

Do you want swag that our awesome marketing team created for this project? Come to our community meetup on prem or contact marketing@sue.nl to receive your swag!

Stickers & Gear

Support

Do you need support that cannot be handled via issue tracking? Please contact us at engineering@sue.nl or via the contact form on this page.

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL