If Site Health shows “Your site could not complete a loopback request” (or The REST API encountered an error) when running WordPress in Docker behind a reverse proxy (e.g., Caddy), the culprit is often that extra_hosts entry pointing to a container IP that changes on restart. Which is always as a solution to solve the loopback issue.
The fix (use a Docker network alias)
Let Docker’s built-in DNS keep things in sync by giving your proxy container a network alias that matches your site’s domain. Then all other containers resolve the alias to the current IP automatically. Would be a better solution than specifying the internal IP for PHP container:
services:
caddy:
# ...your existing caddy config...
networks:
default:
aliases:
- example.comNow remove any extra_hosts lines you previously added to other services (like php).
Apply & verify
# Recreate containers
docker compose up -d
# From the php container, confirm the alias resolves
docker compose exec php getent hosts example.com
# (Optional) Hit wp-cron directly
docker compose exec php curl -I https://example.com/wp-cron.php?doing_wp_cron=1Return to Tools → Site Health and re-check: loopback should pass, and scheduled tasks will run normally.