Outage Lab
Every website students use has someone on call for it at night. When it breaks, that person does not guess: they read the graphs, sign in to the server, find what changed, fix it, check that the fix lasts, and write up what happened so it does not happen again. Outage Lab puts a student in that seat for a made-up school lunch ordering site, with a real Linux terminal, four live graphs and six nights that break in six different ways. It is the work behind every cloud and IT certification, and it plays like a detective game.
Using this with a class? Make a free class and add your students yourself: each gets a username and password, with no email or Google account needed. Or give them one join code. Using Google Classroom? Post the link there; with a free class, the link also puts each student in your class, so you see their work. mytekDev is also available from the Google Workspace Marketplace, so your school's Google admin can install it for everyone.A first lesson: Open Night 1 on the projector and press Acknowledge. Before anyone types, ask the class to read the four graphs out loud: when did it break, and what stopped? Then type uptime. The server says it has been up three minutes. Ask what that means, and let the class choose the next command.
Set up a free class Share to Google ClassroomWhat a student actually does
Each night starts with a phone buzzing on a nightstand. Acknowledging the page opens the console: a status pill, the clock and how many people could not use the site, four graphs (traffic, errors, response time and how busy the machine is), a map of the system with request dots flowing through it, and a terminal signed in to the web server.
The terminal answers about 40 real Linux commands the way an Ubuntu 22.04 server prints them: systemctl and journalctl, tail and grep with pipes into sort and uniq -c, df and du, ss, curl, top, and the tools each night needs, like certbot and ufw. Changing anything needs sudo, as it does on a real server.
Students pin what they find, a log line, a point on a graph or a failing check, and write a theory before they fix anything. A fix that only treats the symptom works for a while: Fast-forward plays the rest of the night, and if the cause is still there, it breaks again and the phone buzzes a second time.
When the site makes it to morning, the student writes an incident report in the order real teams use: summary, impact, root causes, trigger, resolution, detection, action items, lessons, and a step for the next person on call. The timeline fills itself from their night. The report will not print while it blames a person instead of naming the change.
Night 5 at 7:05 a.m.: five times the usual breakfast traffic, and the cache was switched off for a test. Neither turning the cache back on nor adding the second server carries the rush alone; together they do.
What they are learning while they work
Nothing here is presented as a lesson. This is what the activity is made of.
| What the student does | What it is |
|---|---|
| Reads uptime and sees the server has been up three minutes | Evidence first. The fastest way to a cause is what changed. A restart, a deploy, a config edit: the logs and the clock usually say which, before anything is touched. |
| Starts nginx, fast-forwards, and watches it go down again at 4:00 | Symptom and cause. Starting a service fixes tonight. Enabling it fixes every night after. A fix that does not remove the cause comes back. |
| Deletes the 9 GB log and finds the disk still at 100% | Open files. A deleted file a program still has open keeps its space until the program lets go. sudo lsof +L1 shows who is holding it. |
| Restarts the app on Night 3 and nothing changes; rolls back and it works | Rollback. When a release breaks the site, the quickest safe fix is the release before it. The bug gets fixed after, without the clock running. |
| Watches traffic fall on Night 4 while the error rate stays low | The four golden signals. Traffic, errors, latency and saturation each tell a different part of the story. A certificate failure stops browsers before they ask, so it shows in traffic, not errors. |
| Writes Kai forgot and the report will not print | Blameless postmortems. Real teams write up outages without blaming a person, because people stop reporting mistakes when they are blamed. The question is what made the mistake easy to make. |
How to tell whether it landed
Ask, and let them show you. A student who has understood it can answer these without help.
- On Night 1, what is the difference between starting nginx and enabling it?
- Why did deleting the log on Night 2 not free any space?
- On Night 4, why did traffic fall while the error rate stayed low?
- What is the trigger on Night 3, and what are the root causes?
- Night 5 needed two fixes. Why was one not enough?
- Write the step you would leave for the next person on call after Night 6.
Use it in your subject
Not a technology class? Then the app is how students show what they learned in your unit. More projects for every subject.
CSTA 3A-CS-03
Develop guidelines that convey systematic troubleshooting strategies that others can use to identify and fix errors. Every incident report ends with a step for the next person on call.
CompTIA A+ and Linux+
The troubleshooting method in order: identify the problem from logs and recent changes, form a theory, test it, fix it, verify and prevent, document. Linux+ troubleshooting covers services, logs, disk and processes, all of which come up here.
CSTA 3B-NI-03 and 3A-NI-04
Issues that affect network functionality, and scalability and reliability: Night 5's load, a cache and a second server.
Cloud and IT careers
On-call work, monitoring, incident response and postmortems are the daily work behind cloud certifications, site reliability and DevOps roles.
Practical notes
- AgesHigh school IT, CTE, networking and computer science are the home ground. Grade 8 and 9 students can do Nights 1 and 2 with the guide.
- TimeOne period for Nights 1 and 2 and one report. A second period for Nights 3 to 6. The report is a good homework.
- EquipmentA computer or Chromebook with a modern browser. Nothing to install. The server is pretend: no command touches a real machine or the school network. A projector for Big screen.
- CostFree. The app has no paid tier and no AI.
- PrepPlay Night 1 yourself: Acknowledge, sudo systemctl start nginx, Fast-forward, and see it break at 4:00. Then fix it for good. That is the whole demo, about five minutes.
- In a groupGrade level in the header switches between High school (the golden signal names) and Middle school (plain words on the graphs).
For co-ops, microschools, and classrooms
Each night keeps its own state, so a class can play them in any order and come back to one later. With a free account each student's nights and reports are saved as a project, and the shared project shows the graphs around the outage, the stars and the report.
Common questions
Is it free?
Yes. Outage Lab has no paid tier, and it can be tried right on this page without an account. Saving needs a free account.
Is the terminal real?
The terminal is the platform's real terminal, with history and the usual shortcuts, but the server it talks to is pretend. About 40 commands answer the way Ubuntu 22.04 prints them; anything else says command not found. Nothing typed runs anywhere.
What does it simplify?
web-01 is one small machine with a web server, an app server, a database and a cache. A real site has many machines and much longer logs. Time runs at one minute a minute until you fast-forward. The graphs are drawn from the night's story, not measured.
Do students need to know Linux first?
No. The guide and the help command name the first commands, and the first night needs only three. Students who have used the platform's terminal or System Design Lab will recognize everything.
Why does the report refuse to print?
It waits until every section has an answer, the root causes say more than the trigger, and nobody is blamed by name. That is how real postmortems are written.
Are the real outages in it true?
Each night ends with a real outage of the same kind, from the company's own write-up: AWS S3 in 2017, Cloudflare in 2019 and 2025, Fastly in 2021, GitLab in 2017, and a news report of Microsoft Teams in 2020.
Does it need System Design Lab?
No. It stands alone. Students who build servers in System Design Lab will find the same commands and the same kind of logs here, under pressure.
Still have a question? Ask us, and a person will write back.
Where to go next
Start with Outage Lab
Trying it costs nothing and takes about five minutes. An account is what makes the work last.
Outage Lab is one of the making apps on the platform. Every log line, command output and graph comes from the same model, so what a student reads in the terminal is what the graphs show. Nothing in it touches a real server.
Page last reviewed November 2026.
Google, Google Classroom and Chromebook are trademarks of Google LLC.