Delivery / Database migrations / Infrastructure as code

Ship the change. Keep the customers.

A new application wants a renamed field. The old application still reads the original. Release both through a compatibility window and find out whether rollback can actually work.

Try the experiment Learn how it works

The idea in plain language

During a release, old and new application versions may share one database. A compatible migration lets both work until the old readers and writers have been retired.

Actual PostgreSQL executes the migration and client reads. Traffic and readiness phases are an explicit browser model. The download includes a real loopback HTTP switch and a local-only Terraform plan exercise; no cloud deployment is started from this page.

Your independent testing environment

Make the failure visible.

Run the column-rename example. Watch an old client fail after a destructive change. Then compare expansion, backfill and dual writing. The other presets test a failed canary and an early attempt to remove the old column.

Scenario

Compare a destructive rename with expansion, backfill and compatible reads.

Ready. Choose an example and run both implementations.

Explain the result

All safeguarded data reads return 200 across mixed versions and application rollback. The destructive path produces missing-column errors. The contract gate refuses to remove a column while an old reader remains.

Open the exact source behind this edition. The viewer checks each file against its SHA-256 manifest and explains the decision it implements.

Read this extracted package’s setup

Choose your next step

Build the safeguards

Eight explained steps, a source example and a checkpoint at each decision.

Read the tutorial

Compare the alternatives

Understand where each pattern helps and which tradeoffs remain.

Compare approaches

Run and extend it locally

Download the independent environment, run the native checks and try a specific extension.

Open the how-to guide

Before you try it

Do I need to know how to code?

Start with Run and read the comparison. The tutorial introduces the code after you have seen the behavior. The how-to guide separates a no-install route from the native developer exercise.

What happens to the experiment when I reset?

Reset discards this lab’s temporary state and cancels or terminates its pending work. These examples use invented records. No account, production data or cloud deployment is needed.

What does the browser experiment establish?