This exam measures whether you can, on your own, carry a small Kubernetes job through from start to finish: ship an application, update it without taking it down, then repair a failing component. It is meant for people who are finishing level 1, that is, the Fundamentals, Cluster architecture and Basic resources modules. It runs in a virtual machine, on a kind cluster with Kubernetes 1.37, so plan about 60 minutes, and you pass it by meeting the five critical criteria of the grid. It gives no hints and teaches nothing new: it checks what the lessons have taught.
What you will learn
- Say what the exam expects and which state of the cluster it grades.
- Spot the level 1 lessons to review before you start.
- Tell the critical criteria, which alone decide the verdict, from the other two.
- Read your result to decide whether to move on to level 2 or to consolidate.
What should you be able to do before you start?
No step-by-step guide goes with the exam, and choosing the right resources is part of what it measures. Everything it asks is taught in the three level 1 modules: Fundamentals, Cluster architecture and Basic resources. If a step feels unclear, reread those modules before you start.
The Knowledge check page of the Basic resources module is a good entry test: if its questions give you trouble, reread before you attempt the exam.
What does the specification ask for?
The scenario describes an online shop and two situations. It says what must be true at the end, never how to get there: you choose the resources and the commands.
First situation: you must ship the shop's storefront, then update it. In the cluster the storefront is named vitrine.
- The application lives in a namespace named
livraison, which you create, and nothing of it remains in the default namespace. - The storefront runs with three replicas, which replace themselves if one of them disappears.
- It is reachable inside the cluster at
http://vitrine.livraison. - Its welcome message is
MESSAGE_ACCUEIL=Bienvenueand its password isDB_PASSWORD=Boutique-2026. Neither value appears in the image or in the storefront's description: the welcome message is managed separately, and the password is managed as sensitive data, apart from ordinary configuration. - It is shipped first as version 1, then moves to version 2 with no replica missing during the switch, while keeping the ability to go back to version 1.
- The two versions are nginx 1.30.5 then nginx 1.30.6, both pinned by digest:
nginx:1.30.5@sha256:9bf97bd7714f5e24c1ccd545ecb9eb5435cb6d109c97cebb15e7e455e0239edbnginx:1.30.6@sha256:42c68eab6f7e22611786a9fe1d1c019e275c50185bd54b8f75dc2b9184322703
Second situation: the stock component, in the entrepot namespace, no longer starts. It must read a threshold whose value is 10. Before any repair, you write the cause of the failure in ~/epreuve/cause-stock.txt. This file names what is missing (the threshold, its key or its variable) and says that it is missing. The waiting reason that Kubernetes displays is not enough on its own: it says that startup fails, not what is missing. Once the file is written, do not touch it again: its date is compared with the start time of the repaired component. The rule is the following: the component always reads its threshold from the STOCK_SEUIL variable, from a configuration managed separately from its description, and that threshold is still 10 after the repair.
How do you start the exam and get graded?
The lab prepares the cluster and the full statement, then reads the state of the cluster again when you ask for your grade. You start it with dsoxlab run k8s-epreuve-livrer-et-mettre-a-jour. Two commands are then used, and they do not do the same thing. dsoxlab check k8s-epreuve-livrer-et-mettre-a-jour tells you where you stand, criterion by criterion, as many times as you like, but each check acts on the cluster: it launches a temporary Pod, sonde-epreuve, in the livraison namespace to call the storefront, then removes it. dsoxlab submit k8s-epreuve-livrer-et-mettre-a-jour returns the verdict of the exam, applying the critical criteria. The grade never looks at the commands you typed: it reads what is true in the cluster at the moment you ask for it, and the ~/epreuve/cause-stock.txt file for the stock component.
Put it into practice
This lab is the exam itself. It puts you in front of the scenario described above and it is validated on the final state: the grader queries the cluster and reads the ~/epreuve/cause-stock.txt file, as a colleague taking over your work would.
How is the exam graded?
Seven criteria make up the grade, at equal weight: each is worth about 14.3 points out of 100. Five are critical: passing all five is enough to pass the exam, and failing a single one fails it, whatever the rest. The other two, C5 and C6, count toward the grade but do not change the verdict. The lab also sets a pass mark of 70 points, but the five critical criteria are already worth about 71.4 points: that pass mark therefore plays no role, and only the critical criteria decide. A submission can thus total 71 points and fail, because one critical criterion is not met.
| Criterion | Critical | |
|---|---|---|
| C1 | the storefront runs with three available replicas, all on version 2 | yes |
| C2 | the storefront answers at a stable address inside the cluster, served by its three replicas | yes |
| C3 | the configuration and the password live outside the image and outside the storefront's description, with the password managed as sensitive data | yes |
| C4 | an update keeps a way back to version 1 | yes |
| C5 | an update never leaves the storefront without an available replica | no |
| C6 | the application lives in its own namespace, and nothing of it lingers in the default namespace | no |
| C7 | the stock component runs again, and the cause of its failure was recorded before the repair | yes |
What should you do after the exam?
If all the critical criteria pass, level 1 is validated and you can move on to the next level. Otherwise, reread the lesson that teaches the failed criterion, then retake the exam:
- C1: the lesson on deployments;
- C2: the lesson on exposing an application;
- C3: the lessons on configuration and on sensitive data;
- C4 and C5: the lesson on updates and rollbacks;
- C6: the lesson on cluster namespaces;
- C7: the lesson on configuration, which teaches the startup error of a container whose configuration cannot be found.
Key points
- The exam is a measurement: it explains nothing again, gives no hints and grades the final state of the cluster, not your commands.
- The five critical criteria alone decide the verdict: failing one fails the exam, whatever the total.
- The cause of the
stockfailure is written before the repair, and the file is not touched again afterwards. dsoxlab checktells you where you stand as often as you like; onlydsoxlab submitreturns the verdict.