Debugging

Use the failure source to decide how to debug:

If the failure happened during a local deploy, debug from the live local output and local logs.
If the failure comes from GitHub, follow CONTRIBUTING.md and work from the downloaded *job-logs.txt or *.log files.
If a run is still progressing and the user has not asked you to change course, wait for the long-running run to finish instead of interrupting it.

Local Deploy Failures

Before you redeploy, complete all available inspections first. Check the live local output, local logs, and current browser state so the original failure stays visible.
If the website uses locally deployed certificates, run make trust-ca before you inspect it in a browser. Otherwise the browser will warn about the local CA and the inspection will not be reliable.
After make trust-ca, restart the browser so it picks up the updated trust store.
On the first local deploy failure, rerun the affected app with APP=<role> make test-local-dedicated. For alarm failures, use this retry path because it rebuilds the full servers.yml inventory in the dedicated shape.
If you need to validate the single-app servers.yml path after that, continue with APP=<role> make test-local-app.
Use APP=<role> make test-local-rapid only when you intentionally work from the ${TEST_DEPLOY_TYPE}.yml inventory flow created by make test-local-reset or make test-local-full.
Keep iterating locally until the issue is fixed, unless the user explicitly asks you to switch context.

If *job-logs.txt shows a Playwright failure, download the matching Playwright assets before you continue.
Save the downloaded Playwright assets in /tmp/ and keep them there.
Output the storage path for the downloaded Playwright assets file.
Analyze the downloaded Playwright assets together with *job-logs.txt.