Tracking the flow of idea to production for applications deployed to kubernetes.
This is my idea for a potentially useful way to track the 4 Key Metrics (lead time, deployment frequency, mean time to recovery, change failure rate) automatically for applications deployed to k8s.
I'm still hacking on what this should look like. This is currently POC quality, though it can recognize when things are deployed to the cluster.
install docker (yes this is currently only working with docker)
brew update && brew install helm kind helmfile golangci-lint
helm plugin install https://github.com/databus23/helm-diff
sudo sh -c "echo '127.0.0.1 topokube.local' >> /etc/hosts"
make inkind
open https://topokube.local:30443/ui/index.html
make cause-deploy
What does this do? It runs a trivial container, which simulates the deployment of an application. It then makes a request to /metrics to show the current deployment count.
Kubernetes creates a pod for you, which triggers an event that koozie-exporter watches for. Koozie uses the prometheus client to capture metrics. A metrics endpoint (/metrics) is exposed, so that you can scrape these metrics using your own prometheus instance.