Break an integration on purpose and see how it fails

short · 50 min · Objective 2.4

Task

Build a small integration, then induce the three failures that matter -- throttling, pagination and a schema change -- and observe which ones are silent.

Steps

  1. Write lab/integration.py (or equivalent) that fetches a full list and writes the count to lab/integration-count.txt. Do NOT handle pagination yet.
  2. Confirm the count is wrong by fetching the true total another way, and record both numbers in lab/pagination-gap.md.
  3. Fix pagination, and add an assertion that fails loudly if the count changes by more than a defined proportion between runs.
  4. Induce throttling by calling in a tight loop, and record whether your script lost data silently or failed loudly in lab/throttle-result.md.
  5. Change one expected field name to simulate a schema change, run again, and record whether the script degraded quietly or stopped.

Verify

grep -Ec '[0-9]+' lab/pagination-gap.md
grep -Eci 'silent|lost|dropped|loud|failed' lab/throttle-result.md
grep -Eci 'assert|count|threshold' lab/integration.py 2>/dev/null || grep -Eci 'assert|count' lab/integration.*

Both counts recorded and DIFFERENT before the fix, and the throttle result describing which way it failed. The dangerous signature is plausible-but-reduced data, because everything downstream continues to look healthy.

This is an independent study companion for CompTIA SecurityX CAS-005 and is not produced by or endorsed by CompTIA.