Ask HN: How are you doing your smoke testing?

How is everyone doing their smoke testing? My hierarchy of test value basically goes: - Live usage test - Smoke tests - E2E tests - Integration tests - Unit tests And the likelihood of implementation in projects I work on is: - Live usage tests (sanity checks, figuring out if the product flow is what I envisioned) - E2E tests (more commonly for backend projects -- for an API this is basically the only way to ensure quality/functionality/prevent regression) - Integration tests (I get worried about two components actually working together properly) - Unit tests (I generally ignore these, type system does 80% of the work for me) - Smoke tests How is everyone running smoke tests in an automated fashion? All you really "need" is 3 things: - An alert when a staging/production deploy completes - A way to run the test (web-triggerable CI is everywhere now) - (optional) affordances on the app side to set up/teardown state quickly or avoid heavy processes etc. I'm curious what others are doing for this, and what automation/backdoors they put in place. Feels like something I'm overlooking too much these days.