Skip to content

Be the engineer on call.

Something is down. Customers are waiting and the clock is running. Find the real cause, fix it, and do not make it worse. Practise the skill every DevOps and SRE interview asks about, without risking a real system.

PAGE: checkout error rate above 30%

T+0 min
  • CDNHealthy
  • API gatewayDegraded
  • CheckoutDegraded
  • PaymentsDown
  • RedisHealthy
  • PostgreSQLHealthy
  1. 1. You get paged

    A system you have never seen is failing. Customers are hurt and every minute shows on the clock.

  2. 2. Read the evidence

    Open dashboards, logs, deploys and a terminal. Some leads are dead ends, and each one costs time.

  3. 3. Fix it, then explain it

    Pick the fix, name the cause and write what you would change. Some tempting fixes make things worse.

4 outages, each with a different cause

Each one looks the same from the outside and is fixed in a very different way. New cases are added regularly.

  • EasyNew

    Checkout is failing

    Payments start failing at 14:14 on a busy Friday.

    KadePay, a Sri Lankan online shop

    • Kubernetes
    • GitHub
    • PostgreSQL
    • Redis

    About 12 minutes

    Start
  • EasyNew

    The disk that filled itself

    Ride requests start failing at night. Nothing was deployed.

    TukTuk Rides, a three-wheeler booking app

    • Linux
    • Docker
    • AWS
    • GitHub

    About 14 minutes

    Start
  • MediumNew

    Results day, database full

    The results site slows to a crawl as thousands of students log in.

    SmartSchool LK, a school results portal

    • AWS
    • PostgreSQL
    • Docker
    • Nginx

    About 14 minutes

    Start
  • HardNew

    The order that blocks the queue

    Orders are placed but never confirmed. The queue keeps growing.

    Lanka Mart, a grocery delivery app

    • AWS
    • Docker
    • Kubernetes
    • PostgreSQL

    About 15 minutes

    Start

Coming next: a new case lands on Wednesday 25 November. Come back and be on call again.

Why this is the skill that gets people hired

Anyone can follow a tutorial when everything works. Companies hire for the night something breaks. Interviewers ask about the worst outage you handled, and this is how you build an honest answer before you have a job.

  • Evidence before action: the habit that separates calm engineers from frantic ones
  • Blameless post-mortems, written the way real teams write them
  • Real tools in the story: Kubernetes, AWS, Docker, PostgreSQL, GitHub, Linux
  • Nothing real is touched: it is a simulation, safe to get wrong

Planning a certification?

Exam Readiness shows you honestly where you stand and builds a study plan around your exam date.