K8s Goku helps to debug production applications locally without impacting your production system.

Features:

• Relay captured requests to a locally running application (or anywhere the host running the consumer can reach)
• Save requests to disk /when using together with an endpoint this also captures the response and request duraion)
• Transparent to your production traffic
• Multiple consumers can be attached at the same time

Requirements:

• Kuberneters Cluster and the ability to use port-forward
• Project Contour as ingress controller (Others using HTTPProxies might work as well if they support mirroring)
• NodeJS on the host

You need a small deployment to your kubernetes cluster to get the "listener" deployed.

K8s-goku is atm designed to work with Projectcontour.io style HTTPProxies.

kubectl apply -f k8s-deploy/deploy.yaml

Go to edit your HTTPProxy and add this to the services:

- mirror: true
  name: k8s-goku
  port: 8000

Your routes should now look somwthing like this:

routes:
  - conditions:
    - prefix: /
    services:
    - name: my-normal-service
      port: 5000
    - mirror: true
      name: k8s-goku
      port: 8000

Enable a port forward via the Kubernetes API server:

kubectl port-forward k8s-goku-redis-69dbfc9576-x9w5d 6000:6379

This will expose the k8s-goku redis on localhost:6000

To relay the traffic to your local debug app:

node src/consumer.js -r localhost -p 6000 -e <url-of-your-debug-app>

Usage: consumer [options]

Options:
  -V, --version           output the version number
  -e, --endpoint <url>    Endpoint to send requests to
  -n, --name <name>       Name of current run - if omitted a name is generated (run-<timestamp>) (default: "run-1627500598430")
  -s, --save              Save requests and responses
  -r, --redishost <host>  IP/Hostname of redis
  -p, --redisport <port>  port of redis (default: "6379")
  -h, --help              display help for command