Week 17 · lesson

Lesson 4: Mixed Incident Lab

Core path: 42 minutes

Five stations. No domain labels. No trick wording.

Your job is to determine the strongest failure boundary from the evidence, choose the next justified action, and produce a complete support record.

Station 1: hardware/platform fault

User report:

The desktop powers on but never gets to Windows.

Supplied evidence:

  • fans and system power present;
  • monitor and cable verified on another computer;
  • firmware never reaches normal POST completion;
  • board diagnostic indicator points to memory;
  • one DIMM was installed immediately before the failure;
  • no OS change occurred.

Required work:

  1. identify the last proven-good boundary;
  2. explain why Windows repair is not first;
  3. choose the smallest safe hardware check;
  4. state the service/safety boundary;
  5. define success verification.

Station 2: network/service fault

User report:

The internet works, but the class portal will not open by name.

Supplied evidence:

interface: connected
IP configuration: valid
gateway reachability: success
remote test IP: reachable
DNS lookup for portal hostname: fails
other locally cached app content: available

Required work:

  • separate internet path from name-resolution failure;
  • choose the next DNS-related evidence;
  • reject at least one irrelevant physical-network tool;
  • define recovery verification using hostname access.

Station 3: OS/software fault

User report:

My laptop crashes whenever I open the graphics application.

Supplied evidence:

  • Windows boots and remains stable otherwise;
  • other applications run normally;
  • problem began after an application plugin update;
  • application log identifies plugin load failure;
  • previous approved plugin version/configuration is available.

Required work:

  • establish application scope;
  • reject whole-OS reinstall as premature;
  • build the controlled rollback;
  • verify both application startup and the required plugin feature.

Station 4: security/privacy fault

User report:

I keep getting sign-in approval prompts. Can you just turn MFA off?

Supplied evidence:

  • user is not actively signing in;
  • multiple unexpected approval prompts occurred;
  • user did not approve them;
  • endpoint protection is enabled;
  • no supplied evidence proves device malware;
  • organization has an account-security escalation process.

Required work:

  • identify what is observed and what is not proven;
  • preserve the authentication control;
  • choose the safe account-security response/escalation path;
  • explain the answer to the user without blame.

Station 5: operations/change/recovery fault

User report:

The update fixed the problem, but now we need to go back.

Supplied evidence:

  • approved change ticket exists;
  • update solved the original issue;
  • a second business-critical application is now incompatible;
  • rollback package/configuration is documented and available;
  • verified backup was completed before change;
  • maintenance window is still open.

Required work:

  • identify the change-management decision;
  • execute the supplied rollback plan conceptually;
  • verify both the original system state and business-critical application;
  • document outcome and follow-up/escalation.

Support record template

For every station submit:

user complaint:
observed symptom:
scope:
last known change:
safety/privacy/authorization boundary:
last proven-good layer:
first failed or unproven layer:
strongest theory:
rejected alternative:
best evidence/tool:
controlled action:
rollback/recovery:
verification:
customer-facing explanation:
documentation/escalation:

Peer review

Exchange one completed station with a partner.

The reviewer must answer:

  • Does the evidence actually support the theory?
  • Is the proposed action smaller than a wipe/reset/reinstall when a narrower fix exists?
  • Is rollback/recovery realistic?
  • Does verification test the required function?
  • Did the technician respect safety, privacy, and authorization?

Success criteria

You pass when all five incident records are internally consistent and a reviewer can follow the reasoning without needing the domain label.

Read it. Prove it.

Lesson knowledge checks

Answer from the lesson you just completed. Results stay in this browser and are not submitted.
Knowledge check 1

A mixed incident includes slow Wi-Fi and a printer paper jam. What is the correct approach?

Knowledge check 2

Why are deterministic fault stations useful in training?