Show HN: A playable version of the Claude Code Terraform destroy incident

YouBrokeProd - Production Incident Simulator for SREs & DevOps Engineers ● PagerDuty is screaming. Your phone won't stop. It's 3 AM. Practice On-Call Before the Real Page Comes In Hands-on production incident simulations for SREs, DevOps engineers, and technical founders. A real-time game simulation you play, laugh, and learn from. Drop into a realistic terminal with a ticking clock and a system on fire. Run commands, find the root cause, and fix it before time runs out. No setup required. Play Now - No Signup Required or sign up for 10 scenarios, leaderboard, and more Runs in your browser. Takes 5 minutes. The only production incident simulator you can start in seconds. 279 + incidents simulated 118 + registered engineers Like It? There's 9 More. Sign up free with GitHub or Google - takes 10 seconds ✓ The viral Terraform scenario (685K+ views) ✓ 9 more incidents across databases, Kubernetes, cloud, and security ✓ Leaderboard ranking against other engineers ✓ Score breakdown and solution walkthroughs Sign Up Free - 10 Seconds Trending 685K+ views - covered by Tom's Hardware, Hacker News, and more An AI agent ran terraform destroy on production. 2.5 years of data - gone. Can you recover it? Based on the real DataTalksClub incident that hit the front page of Hacker News. Play through an authentic Claude Code split-panel interface - the same kind of setup that caused the original disaster. Sign Up to Play Free account required - 10 seconds with GitHub Play This Scenario What the DevOps community said when this happened in real life AG Alexey Grigorev @Al_Grigor "Claude Code wiped our production database with a Terraform command. It took down the DataTalksClub course platform and 2.5 years of submissions." 685K+ views CE Christoph Engelbert @noctarius2k "Just as someone posts that Claude Code deleted a production environment via Terraform, we see 'all those annoying manual approvals need to go away'" Y Hacker News Front Page "I call BS on anyone who says they check every little thing their agent does. This will happen more, not less." 500+ comments Not a Tutorial. Not a Quiz. A Real-Time Incident Simulation. You get dropped into a live production incident with real tools. No hand-holding. Figure it out before the clock runs out. youbrokeprod.com/play/db-connection-pool-001 Try It Now - Free >_ Real Terminal Run actual commands - kubectl get pods , tail -f logs , db connections . Tab completion, command history, the works. 📊 Live Logs and Metrics Streaming log panels, error spikes, connection pool gauges - the same signals you'd see in Datadog or Grafana during a real incident. ⏱️ Ticking Clock Every scenario has a par time. Revenue ticks down. PagerDuty alerts keep firing. The pressure is the point - it trains you to think clearly under stress. 🔍 Diagnose the Root Cause Gather clues, eliminate red herrings, then commit to your diagnosis. Wrong guess? Points deducted. Right answer unlocks the fix phase. 📝 Post-Incident Debrief Every scenario ends with a full breakdown - root cause, optimal investigation path, what you missed, and the fix explained. Learn from every run. ☕ Power-Ups and Superpowers Coffee for bonus time. Deep Focus mode to block distractions. Hints when you're stuck. Use them wisely - they cost points. Try It Free - No Signup Required Your first production incident shouldn't be your worst one. Most engineers and technical founders get paged cold with zero prior experience handling a real incident. Reading runbooks doesn't build on-call instincts. YouBrokeProd drops you into realistic incident simulations so when the real page comes in at 3 AM, you've already been there. On-Call Skills That Actually Stick 10 incident scenarios across beginner, intermediate, and advanced difficulty 🗄️ Triage Database Failures Fast Read Postgres error states, diagnose connection pool saturation, and fix replication issues without guessing. ☸️ Debug Kubernetes Under Pressure Diagnose crashloops, OOMKills, and networking failures systematically instead of opening Stack Overflow. 🔒 Spot Security Issues Before They Escalate Recognize credential exposure patterns, suspicious traffic, and misconfigurations that lead to real breaches. How It Works Each scenario is a real-time simulation running in your browser. No setup. Just you, a terminal, and a production incident to solve. 1 Get Paged Pick a scenario and difficulty. You get a briefing with symptoms, a simulated terminal, and a ticking clock. 2 Investigate Run real commands in the terminal - check logs, query metrics, inspect configs. Built-in hints if you get stuck. 3 Diagnose & Fix Submit your root cause diagnosis, then apply the fix command. Scored on speed, accuracy, and efficiency. 4 Debrief See what you got right, what you missed, and the optimal diagnostic path. Compare your score on the leaderboard. On-call training for your whole team? Run the same incident simulation across your SRE, platform, or founding engineering team. Compare scores, identify skill gaps, reduce MTTR, and build shared muscle memory for when the real pages come in. Manager reports and team leaderboards included. See Team Plans The Next Incident Won't Wait. Will You Be Ready? Start your first incident simulation right now - no signup, no credit card, no setup. Start Your First Simulation