Skip to content

C7: Diagnostics

  • Do now: Turn everything you know about FreshConnect into a one-page incident card.
  • Input: Platform knowledge from C2-C6, plus any deployed resources and docs.
  • Output: agent-output/freshconnect/07-diagnostics-quick-card.md.
  • Required to move on: Top health checks, symptom-to-action flows, key commands or queries, and an escalation trigger.
  • Decisions now: What to check in the first 60 seconds, what failures are most likely, and when to wake up someone else.
  • Next: C8 uses this card as evidence that your solution is operable under pressure.

The goal is not a full runbook. It is a fast triage aid that works at 2 AM when the reader has no spare attention.

At 2:17 AM on a Sunday, the FreshConnect API P95 latency spikes to 8.2 seconds and the on-call engineer is new to the rotation. They need a triage card that tells them what to check first, what symptoms mean, and when the problem is bigger than a quick fix.

  1. Pick the top three health checks that should happen in the first 60 seconds.
  2. Add two or three symptom-to-action flows for the failures you think are most likely.
  3. Include the key Azure CLI commands or KQL queries that make the card actionable.
  4. Define the escalation trigger and save the final card at the required path.
  • Which checks tell you fastest whether the issue is platform, application, or dependency?
  • Which failure modes matter most for FreshConnect based on your architecture and load-test results?
  • What can an on-call engineer do safely without architect approval?
  • How do you keep the card short enough to scan while still being useful?
  • agent-output/freshconnect/07-diagnostics-quick-card.md
  • Top 3 health checks for the first minute of an incident.
  • Two or three symptom-to-action flows.
  • Key commands or queries for diagnosis.
  • Escalation trigger and optional path to agent-output/freshconnect/07-ab-diagnostics-runbook.md if you add bonus depth.
FocusWhat good looks likeEvidence
First-minute triageThe card helps someone act immediatelyThree health checks are specific and ordered
Symptom handlingCommon failures map to concrete next actionsTwo or more symptom-to-action flows are included
ActionabilityThe card is usable without extra explanationCommands, queries, or portal checks are present
Escalation disciplineThe card defines the handoff clearlyClear trigger or threshold tells on-call when to escalate
What belongs on the card

Prioritize checks that answer “is it down, slow, or broken by a dependency?” Use Hints & Tips for example diagnostic flows and Quick Reference Card for example queries.

  • A diagnostics card full of paragraphs will not help during an incident.
  • Vague escalation language creates unnecessary delay.
  • Do not include five different branches if only two or three are likely to matter.
  • Avoid commands that require hidden context the on-call engineer will not have.
ItemValue
Input fromDeployed infrastructure, prior artifacts (architecture, IaC templates, ADR)
Your outputagent-output/freshconnect/07-diagnostics-quick-card.md (required), optionally agent-output/freshconnect/07-ab-diagnostics-runbook.md
Next challenge usesC8 uses this card as evidence that your solution is operable, not just deployable

Challenge 8 uses this quick card alongside your diagram, ADR, and other artifacts to show operational maturity during the final presentation.