Exercise 3: 5 Whys — "Images Are Loading Slowly" — Possible Solution
====================================================================

Starting symptom: "Users report the website is loading images
slowly."

WHY? -- Because image requests are taking much longer than usual to
respond.

WHY? -- Because the object storage bucket serving the images is being
accessed directly from the application's origin server on every
request, rather than through a caching layer, and that origin path has
gotten noticeably slower recently.

WHY? -- Because a recent traffic increase pushed the origin server's
requests-per-second against the object storage service's per-bucket
request rate limits, causing a portion of image requests to be
throttled (Ch.4/Ch.5's storage-tier and throttling material).

WHY? -- Because the site was originally built to serve images directly
from the storage bucket without a CDN or edge cache in front of it, an
architecture choice that was fine at the site's original, much lower
traffic volume but was never revisited as traffic grew.

GENUINE ROOT CAUSE (not just a first-level symptom): the architecture
never added a caching/CDN layer in front of object storage as traffic
scaled up, meaning every single image request -- not just a handful --
depends directly on the storage service's own request-rate limits,
which is now the actual bottleneck under current traffic levels.

Why this is a genuine root cause rather than just a deeper-sounding
symptom: fixing this (adding a CDN/caching layer in front of the image
storage) would prevent the SAME underlying problem from recurring as
traffic continues to grow, because it removes the architectural reason
every image request depends directly on a rate-limited service in the
first place. Merely raising the storage service's rate limit (a
shallower fix, addressing only the third "why") would likely provide
temporary relief but leave the site's IMAGE-SERVING PATTERN itself
unchanged, meaning the same problem would likely resurface again once
traffic grows further -- exactly the "fixes the symptom, not the root
cause" trap this chapter's own warn-box describes.

WHY THIS WORKS AS AN ANSWER
------------------------------
Each "why" moves one level deeper than the last, following the
chapter's own worked example structure (slow site -> database slow ->
pool exhausted -> deployment introduced a leak -> no automated test
caught it), ending on an architectural/process-level explanation
rather than stopping at a technical symptom like "storage was
throttled" -- exactly the distinction the chapter draws between a
genuine root cause and a merely deeper-sounding restatement of the
same surface problem.