Amplify Diagnostics

AWS Amplify Hosting Diagnostics

When to use

Any AWS Amplify Hosting investigation where the console alone is insufficient — build failures, deployment issues, custom domain configuration, SSR problems, backend integration, access control, or CDN performance troubleshooting.

Investigation workflow

Step 1 — Collect and triage

aws amplify list-apps --query 'apps[*].{Id:appId,Name:name,Platform:platform,Repository:repository}'
version: "1.0.0"
last_updated: "2025-04-12"
aws amplify get-app --app-id <app-id>
aws amplify list-branches --app-id <app-id> --query 'branches[*].{Name:branchName,Stage:stage,Status:activeJobSummary.status}'
version: "1.0.0"
last_updated: "2025-04-12"
aws amplify list-jobs --app-id <app-id> --branch-name <branch> --max-results 10

Step 2 — Domain deep dive

aws amplify get-job --app-id <app-id> --branch-name <branch> --job-id <job-id>
aws amplify get-branch --app-id <app-id> --branch-name <branch>
aws amplify list-domain-associations --app-id <app-id>
aws amplify get-domain-association --app-id <app-id> --domain-name <domain>

Step 3 — Detailed investigation

aws amplify list-backend-environments --app-id <app-id>
aws amplify get-backend-environment --app-id <app-id> --environment-name <env>
aws cloudfront list-distributions --query 'DistributionList.Items[?Comment==`<app-id>`]'
aws logs describe-log-groups --log-group-name-prefix /aws/amplify/<app-id>

Read references/amplify-guardrails.md before concluding on any Amplify issue.

Tool quick reference

Gotchas: AWS Amplify Hosting

• Build settings can come from amplify.yml OR console. The amplify.yml file in the repository root takes precedence over console build settings. If both exist, the file wins. This is a common source of confusion when console changes don't take effect.
• Environment variables have scope. App-level variables apply to all branches. Branch-level variables override app-level. Some variables are reserved (e.g., _LIVE_UPDATES). Secrets are not exposed in build logs.
• Custom domains require DNS verification. Amplify creates CNAME records for verification. DNS propagation can take up to 48 hours. SSL certificates are automatically provisioned via ACM. Subdomains and root domains have different configuration requirements.
• SSR requires Amplify Hosting compute. SSR apps use Lambda@Edge or CloudFront Functions. Compute settings affect cold start and memory. Not all frameworks are supported for SSR. Build output must include server-side artifacts.
• PR previews create temporary deployments. They use a subdomain pattern. They are automatically deleted when the PR is closed. They share the app's environment variables unless overridden. They can increase build minutes usage.
• Rewrite and redirect rules have order precedence. Rules are evaluated top to bottom. The first matching rule wins. Trailing slashes matter. Regex patterns are supported but must be valid.
• Branch auto-detection connects to the repository. It requires repository access permissions. New branches are automatically deployed. Pattern matching controls which branches are deployed. Disconnecting the repository stops auto-detection.
• Basic auth protects the entire branch. It cannot be applied to specific paths. Credentials are branch-specific. It uses HTTP Basic Authentication. It is not a substitute for application-level auth.
• Amplify uses CloudFront for CDN. Cache invalidation happens automatically on deploy. Custom cache headers can be set in custom headers configuration. CloudFront distribution settings are managed by Amplify.

Anti-hallucination rules

1. Always cite specific app IDs, job IDs, or API responses as evidence.
2. amplify.yml takes precedence over console build settings. Never assume console settings apply when the file exists.
3. Custom domains require DNS verification. Never assume instant domain activation.
4. SSR requires compute configuration. Never assume static hosting supports SSR.
5. Environment variables have scope precedence. Never assume app-level vars override branch-level.
6. Spend no more than 2 minutes on any single hypothesis. Pivot if inconclusive.

22 runbooks