Challenge 3: Diagnose a Connection Failure — Possible Solution
====================================================================

MOST LIKELY CAUSE: web and db are attached to DIFFERENT Docker
networks (or one is on the default bridge and the other is on a
custom network) — meaning there is genuinely no network path between
them at all, regardless of whether both containers are individually
"running" without any errors.

This matches this chapter's own gotcha directly: "both running" only
confirms each container's PROCESS started successfully — it says
nothing about whether the two containers share a network that would
let them reach each other. This isn't a firewall issue, a permissions
issue, or an application-level bug; it's Docker's network isolation
behaving exactly as designed, just not configured the way the
developer assumed.

HOW TO CONFIRM: run docker network inspect on the network(s) each
container is attached to (or check docker network ls / the relevant
Compose file's networks: configuration) and verify whether web and db
actually share a common network. If they don't, the fix is either
attaching both to the same network, or — if they were deliberately
isolated on purpose — routing the connection through an intermediary
service (like this chapter's api example) that IS attached to both.