EnlightLab
View staging app Start guided demo →
Production self-healing for Kubernetes

Deploy, break, explain, and recover — live on Kubernetes

A client-ready demo: deploy a FastAPI app with GitOps, simulate a real outage, let AI explain what failed, then restore with one click. Integrates with ArgoCD, k8sgpt, and Slack for your team's workflows.

View live staging app → Start guided demo →

① Deploy

Start from zero — deploy FastAPI to staging via ArgoCD and verify health checks.

② Break

Inject a bad image safely. ArgoCD shows Progressing → Degraded while the app fails.

③ Explain

AI reads pods, events, and cluster state — plain English, not ConfigMap noise.

④ Heal

GitOps rollback restores the good image, re-syncs ArgoCD, and verifies recovery.

Architecture

What runs where

Everything runs on a managed Kubernetes cluster (Mumbai) — real VMs, real cluster, not slides.

GitHub Manifests & GitOps
Argo CD Sync & deploy
Kubernetes Mumbai region
Kubernetes cluster — 3 namespaces
selfheal

Demo control UI · kubectl + k8sgpt

enlight-staging

FastAPI workload pods

argocd

GitOps controller

Container registry enlight-fastapi images
FastAPI image enlight-fastapi
Health checks /health endpoint
Load Balancer · selfheal-ui
/demoGuided wizard /staging/Live FastAPI app
Slack integration

Production deployments notify your team in Slack on deploy, failure, and auto-heal — same pipeline pattern for Flask, Node.js, and custom stacks.

See live workflow →