The heroku surrogate command pulls down the environment variables for a
Heroku application, merges them into the local environment, and spawns a
process. Here's a Rails specific example:
heroku surrogate rake db:migrate:up VERSION=20130228193803
This runs the migration locally, but connected to the DATABASE_URL specified
by the server. This is great for sneaking a nondestructive migration out
before the actual deploy, thus minimizing downtime.
The fun doesn't stop there. You can pass in a process type from your
Procfile or any arbitrary command. Spin up a console. Spin up a worker.
Heck, spin up a web server if you want.
heroku surrogate console
heroku surrogate worker
heroku surrogate PORT=3000 web
The --release option lets you use the environment and process types from an
earlier release.
heroku surrogate --release=v123 env
Note that you're using development code against production resources. If
you'd rather be running production code as well, pass --checkout, and the
appropriate commit (respecting --release if given) will be checked out
before executing.
heroku surrogate --checkout foreman start
heroku plugins:install https://github.com/tpope/heroku-surrogate.git