atlas-migrations

Overview

Atlas manages and migrates database schemas as code - define the desired schema and Atlas plans, lints, tests, and applies the changes (per atlasgo.io/getting-started (opens in new window)). Two operating modes:

1. Declarative - define desired state in HCL or SQL; Atlas computes diff and applies (atlas schema apply).
2. Versioned - Atlas generates timestamped migration files from schema diffs (atlas migrate diff); migrations apply forward (atlas migrate apply).

Most production teams use versioned mode for the audit trail; dev loops often use declarative mode for fast iteration.

How to use

Follow Steps 1-9 in order - lint (Step 6) is the gate before apply (Step 5).

Step 1 - Install

Per at-start (opens in new window):

# download the install script, then run it
curl -sSf https://atlasgo.sh -o install-atlas.sh
sh install-atlas.sh

# Homebrew
brew install ariga/tap/atlas

# Direct binary download (per platform):
#   atlasbinaries.com/atlas/atlas-{linux-amd64,linux-arm64,
#                                  darwin-amd64,darwin-arm64,
#                                  windows-amd64}-latest

Step 2 - Define schema

Per at-start (opens in new window), schema can be SQL (most common) or HCL. SQL example (PostgreSQL):

-- schema.sql
CREATE TABLE users (
  id serial PRIMARY KEY,
  name varchar(255) NOT NULL,
  email varchar(255) UNIQUE NOT NULL,
  created_at timestamptz DEFAULT now()
);

For the same schema expressed in Atlas's HCL DSL and guidance on when HCL vs SQL matters, see references/hcl-schema.md.

The dev DB referenced via --dev-url is a temporary scratch database Atlas uses to compute the diff (Atlas applies the desired state to the scratch DB, then computes the migration vs production).

Step 3 - Apply schema (declarative mode)

Per at-start (opens in new window):

atlas schema apply --url "$DATABASE_URL" --to file://schema.sql \
  --dev-url "docker://postgres/17/dev?search_path=public"

--url = target DB; --to = desired schema definition; --dev-url = scratch DB for diff computation. Atlas prints the planned changes and prompts for confirmation by default.

Step 4 - Generate versioned migrations

Per at-start (opens in new window):

atlas migrate diff initial --to file://schema.sql \
  --dev-url "docker://postgres/17/dev?search_path=public"

This generates a timestamped migration file in migrations/ (e.g., 20260506120000_initial.sql). Subsequent calls with a new schema generate new migrations against the previous state.

Step 5 - Apply versioned migrations

Verify (gate before applying): run Step 6 lint first and assert it reports no destructive / data-loss findings; if it flags one, fix the migration (add a DEFAULT, split the change, or add a new migration) and re-run lint until clean. Only then apply:

atlas migrate apply --url "$DATABASE_URL"

Atlas tracks state in atlas_schema_revisions table - each applied migration is recorded with its hash, so tampering is detected on subsequent applies.

Step 6 - Lint migrations

Per at-start (opens in new window):

atlas migrate lint --dev-url "..." --latest 1

Lints the most recent migration for: data loss, narrow-column operations (e.g., varchar(50) → varchar(20)), missing default on a NOT NULL add, lock-escalating ops on large tables, schema backwards-incompatibility.

The lint output is the value-add over Flyway / Liquibase - those tools are syntax-only; Atlas knows about destructive patterns.

Step 7 - Hash + integrity

Atlas computes hashes for each migration file. If a developer edits an already-applied migration, atlas migrate apply fails until atlas migrate hash re-syncs (intentional act). This protects against silent migration tampering.

Step 8 - CI integration

- uses: ariga/setup-atlas@v0
- name: Lint migrations
  run: atlas migrate lint --dir "file://migrations" \
    --dev-url "docker://postgres/17/dev?search_path=public" \
    --latest 1
- name: Apply migrations to staging
  run: atlas migrate apply --url "${{ secrets.STAGING_DB_URL }}"

GitHub Action ariga/setup-atlas@v0 provides Atlas in the runner; the atlas-action orchestrator provides PR-comment integration with diff visualization.

Step 9 - Composition with sister tools

Beyond atlas migrate lint, apply adversarial review with team-specific risk policies that Atlas's built-in lint doesn't capture (e.g., "no DROP TABLE without DBA approval").

Worked example

Add a NOT NULL column status to a 40M-row users table:

1. Edit schema.sql to add status varchar(20) NOT NULL DEFAULT 'active' on users.
2. atlas migrate diff add_status --to file://schema.sql --dev-url "docker://postgres/17/dev?search_path=public" writes migrations/20260506120000_add_status.sql.
3. atlas migrate lint --dev-url "docker://postgres/17/dev?search_path=public" --latest 1 passes because the NOT NULL add ships a DEFAULT (a NOT NULL add with no default is flagged).
4. CI runs the same lint on the PR, then atlas migrate apply --url "$STAGING_DB_URL" records the migration in atlas_schema_revisions.
5. A teammate later edits that applied file; their atlas migrate apply fails on hash mismatch until an intentional atlas migrate hash re-sync - the tamper guard working as designed.

Anti-patterns and limitations

The failure modes (editing applied migrations, skipping --dev-url, declarative schema apply straight to production, skipping lint in CI, blind atlas migrate hash) and the tool's limitations (partial HCL for some DBMS features, general-purpose lint rules, --dev-url requiring a real DBMS, paid Atlas Cloud) are catalogued in references/atlas-caveats.md.

References

• at-start (opens in new window) - getting-started, install commands, example schema, all atlas commands, supported databases
• atlasgo.io/docs - full documentation
• github.com/ariga/atlas - repository
• flyway-migrations, liquibase-migrations, sqlmesh-migrations - sister tools (Flyway / Liquibase = imperative versioned; SQLMesh = data-pipeline + schema)