Connect Postgres for Sinatra · Ample Computer

Connect Postgres for Sinatra

Database connection for a Sinatra app on Ample. Verified on Sinatra: a managed PostgreSQL database auto-provisioned and injected as DATABASE_URL, with real queries succeeding at request time (records=ok). Build and start: bundle install into vendor/bundle from Gemfile.lock (development and test groups skipped), then Puma via rackup (config.ru) reading PORT on the ruby-3.4 template.

Verified November 5, 2026

What it runs on

You need

Exact tested configuration

installbundle config set --local path vendor/bundle && bundle config set --local without 'development test' && bundle install

runtimeruby

sizes-1vcpu-1gb

startbundle exec rackup -o 0.0.0.0 -p $PORT

templateruby-3.4

Steps

  1. Build and start. bundle install into vendor/bundle from Gemfile.lock (development and test groups skipped), then Puma via rackup (config.ru) reading PORT on the ruby-3.4 template; the server must bind 0.0.0.0 on PORT.

  2. Read DATABASE_URL at runtime. Open connections from the environment inside handlers or a pool; omit DATABASE_URL to auto-provision a managed database or pass --env DATABASE_URL=... for an existing one.

  3. Verify. Fetch the live URL and /p/relational-records on the example; on failure read the build and runtime logs.

$ ample logs <deployment_id> --kind build

How it was checked

✓ app responds on its public URL· http get/ → “ample canary sinatra patterns”
✓ relational-records check from the example· http get/p/relational-records → “see the pattern fixture checks”

Evidence

  1. November 5, 2026 · canary run

Sinatra pattern fixture deployed on Ample (bundle install into vendor/bundle from Gemfile.lock (development and test groups skipped), then Puma via rackup (config.ru) reading PORT on the ruby-3.4 template); for this guide: a managed PostgreSQL database auto-provisioned and injected as DATABASE_URL, with real queries succeeding at request time (records=ok).

Tested examples

What it doesn’t cover