Smoke Test Runner

Overview

Execute fast, high-confidence smoke tests that validate critical application functionality after deployment or build. Smoke tests verify that the application starts, core user flows work, and key integrations respond -- without running the full test suite.

Prerequisites

• Application deployed and accessible at a known URL or running locally
• HTTP client available (curl, wget, node-fetch, or Playwright)
• List of critical endpoints and user flows to validate
• Expected response codes and content patterns for each check
• CI/CD pipeline hook for post-deployment validation

Instructions

1. Identify the critical paths that constitute a "working" application:

• Health check endpoint returns 200 with expected body.
• Homepage loads and contains key UI elements.
• Authentication flow succeeds with test credentials.
• Primary API endpoint returns valid data.
• Database connection is active and responding.

1. Create a smoke test configuration listing each check:

• URL or command to execute.
• Expected HTTP status code (200, 301, etc.).
• Response body pattern to match (substring or regex).
• Maximum acceptable response time (e.g., 3 seconds).

1. Write the smoke test suite as a lightweight script or test file:

• Use curl for HTTP checks or Playwright for browser-based checks.
• Run checks sequentially for simplicity (parallel for speed if independent).
• Fail fast on the first critical failure.
• Log each check result with pass/fail, response time, and status code.

1. Implement timeout guards:

• Set a global timeout of 60 seconds for the entire smoke suite.
• Set per-check timeouts of 5-10 seconds.
• Treat timeouts as failures, not retries.

1. Add deployment-gate integration:

• On success: proceed with deployment promotion or traffic shifting.
• On failure: trigger rollback and send alert notification.
• Report results to CI/CD dashboard and Slack/Teams webhook.

1. Store smoke test results as CI artifacts for audit trail.
2. Schedule periodic smoke runs (every 5 minutes in production) as synthetic monitoring.

Output

• Smoke test script (scripts/smoke-test.sh or tests/smoke.test.ts)
• Pass/fail result for each critical check with response times
• Deployment gate verdict (PASS or FAIL with reason)
• CI artifact with timestamped smoke test log
• Alert payload for failed checks (Slack webhook, PagerDuty, etc.)

Error Handling