IndustriesSoftware & IT operations
Practise handling outages before your first real one
The website is down and customers can't pay. Find out why, fix it safely and keep everyone informed, in a simulated company where nothing real breaks.
- Runs in
- A web browser, 3D included
- One exercise
- 20 to 55 minutes
- Practice levels
- Six, from worked example to exam
- Working tools
- 4 used in these exercises
A real exercise
Checkout is failing for a third of customers. Alerts are firing, the support team is asking what to tell people, and a release went out an hour ago. You open the incident console, trace the errors to the new release, switch off the feature that broke it, and write the update for the business.
How it works
Every exercise follows the same rhythm as the real job: a problem arrives, you investigate, you act, and the simulated company responds.
- 01
Something breaks
An alert, a customer complaint, or a security warning starts the clock.
- 02
You investigate
Read dashboards, logs, the release pipeline and the alert queue. Ask the engineers who own the service.
- 03
You act
Roll back, switch off a feature, contain a compromised machine, and post updates people can understand.
- 04
You see the impact
Lost orders, time to recovery and what you told whom are all on the record for the review afterwards.
What you practise
Concrete tasks taken from the exercises on this page. Each one is done with real data, under real constraints, inside the same simulated company.
- T01
Lead an incident
Take charge of an outage, decide what matters first, and bring the service back.
- T02
Find a bad release
Follow the trail from the errors to the change that caused them.
- T03
Use feature flags safely
Switch off what broke without switching off what works.
- T04
Respond to a security alert
Triage the alert, contain the affected system and preserve evidence.
- T05
Communicate during a crisis
Write clear updates for managers, support staff and customers while the work goes on.
What your organisation gets
Evidence, not attendance.
- 01
On-call confidence
New engineers face their first outage in practice, not at 3 a.m. with real customers waiting.
- 02
Drills without risk
Run incident drills without touching production systems or staging an outage.
- 03
Reviews based on facts
Every action and message is recorded, so the review discusses what really happened.
The tools you work with
Working tools, not screenshots. They read the same live data as the rest of the simulated company, and what you change in one shows up in the others.
- 01
Incident console
The live state of services and hosts during an outage, with the runbook actions.
- 02
Feature flags
The switches a service reads, and which one is breaking it.
- 03
Data pipelines
The pipelines feeding production, and which one is sending bad data.
- 04
Security alerts
Alerts, accounts, endpoints and signs of compromise in one view.
Six levels, one exercise
Teachers and trainers choose how much help each person gets. The exercise stays the same; only the guidance and the grading change.
- L1
Watch it done
The exercise solved step by step, with what each step changed.
- Hints
- Read only
- Graded
- No
- L2
Step by step
The next move is pointed out, every hint is open, goals update as you work.
- Hints
- All hints
- Graded
- Yes
- L3
With hints
Every hint is open, but finding the next step is up to you.
- Hints
- All hints
- Graded
- Yes
- L4
On your own
Hints stop short of the worked answer. Goals still update as you go.
- Hints
- Some hints
- Graded
- Yes
- L5
Exam
No hints and no AI helper. Results appear only after you submit.
- Hints
- None
- Graded
- Yes
- L6
Free play
An open what-if sandbox. Try anything; nothing counts towards a grade.
- Hints
- All hints
- Graded
- No
What every attempt leaves behind
Judge the work, not a quiz score. Every attempt is recorded as it happens, so you can see how someone worked, not only where they ended up.
Every action, in order
What was checked, who was asked, what was changed and when. Nothing is overwritten.
Scores by skill
Each exercise scores the skills it was built to test, from what was actually done.
Weak habits, flagged
Poor patterns in how the work was done are flagged, not only wrong final answers.
A replay of the attempt
Step through the session as it happened, with typed personal details masked.
- Something breaks
- You investigate
- You act
- You see the impact
Do we connect our real systems?
No. Everything happens inside a simulated company with its own services, alerts and people, so nothing real can break.
Who is this for?
Students of computer science and IT, new engineers joining an on-call rotation, and teams that want to rehearse incidents.
Is there AI help?
An AI helper can give hints during practice. It is switched off in assessments, and organisations can turn it off completely.
Related
See it with your own people
Start with a pilot: one course or one team. We help you choose the exercises and measure the results.