Skip to main content

Flyway Snapshot Alternative in Atlas

The flyway snapshot command captures the structure of a database and stores it as a JSON file that represents its state at a specific point in time. This is often used for drift detection, change reviews, or as a baseline for further comparisons.

In Atlas, you can achieve the same workflow with a few key advantages using the atlas schema inspect command.

Inspect and Snapshot Your Schema​

Unlike Flyway's snapshot feature, this capability is free in Atlas. With the schema inspect command, you can capture your database schema and export it in any format you need. The examples below show how to export a database to a SQL (DDL format), HCL, JSON, or even a Mermaid diagram. Next sections show more advanced exports, such as generating ORM models or structured code folders from your database schema.

atlas schema inspect -u "<DATABASE_URL>" --format '{{ sql . }}' > schema.sql

This command connects to your database, inspects its structure, and produces a complete, version-controlled representation of your schema. The output can be stored in your Git repository, compared against other states, or used to detect drift between environments.

What the Snapshot Includes​

Atlas currently supports introspection for the following objects in the free version:

  • Schemas, tables, and columns
  • Indexes and unique constraints
  • Foreign keys and relationships
  • Check constraints
  • Enums (PostgreSQL, MySQL, etc.)
  • Comments

These are the core elements required for most schema-as-code and drift-detection workflows.

Using Atlas Pro, you can extend the schema snapshot to provide the full representation of your database's state, with complete dependency-graph and relationships between the different objects, including:

  • Views and materialized views
  • Functions and procedures
  • Triggers
  • Sequences
  • Custom types, such as domains, composite-types, etc.
  • Extensions, event-triggers, RLS policies, and more.

Customizing Snapshot Output​

Atlas supports exporting the schema to custom formats using Go templates. This allows you to generate tailored outputs for your specific needs. For example, you can create documentation, custom reports, or export ORM definitions. For example:

📺 For a step-by-step example walk-through, watch our tutorial: Inspect Your Database Schema with Atlas

Visualize with ERD​

You can also visualize your schema as an interactive Entity Relationship Diagram (ERD). Use the --web flag to open an interactive ERD in Atlas Cloud:

atlas schema inspect -u "<DATABASE_URL>" --web

Try it out in the component below:

Loading ERD...

Export Database Schema to Code​

Users who want to export the database to structured folders can follow the Export Schema to Code documentation. Atlas can organize your schema into a modular directory structure like this:

├── extensions
│   ├── hstore.sql
│   └── citext.sql
├── schemas
│   └── public
│       ├── public.sql
│       ├── tables
│       │   ├── profiles.sql
│       │   └── users.sql
│       ├── functions
│       └── types
└── main.sql

To export your schema into this structure, use the following command:

atlas schema inspect -u "<DATABASE_URL>" \
  --format '{{ sql . | split | write }}'

This will create a structured directory with your schema organized by object type (tables, views, functions, etc.), with a main.sql file as an entry point.

Comparing Snapshots (Drift Detection)​

Once you have a snapshot of your schema, you can compare it to another state - a live database, a migration directory, or another snapshot file.

atlas schema diff \
  --from "file://schema.hcl" \
  --to "postgres://localhost:5432/prod?sslmode=disable"

This command produces a deterministic diff between two schema states, allowing you to identify and review changes before applying them.

Learn more about drift detection and comparing schemas.

Use Cases Covered (Compared to Flyway)​

Use CaseFlyway CommandAtlas EquivalentNotes
Snapshot live databaseflyway snapshot -source=<db> (Enterprise)atlas schema inspect -u <url>Atlas allows exporting to SQL, HCL, JSON, custom templates, etc.
Drift detection between snapshotsflyway checkDriftatlas schema diffAtlas supports comparing any sources
Advanced filtering-source / -schemaModelSchemas--schema, --include, --excludeFlyway is limited to filter by schemas; Atlas allows filtering by schemas, type, name, glob, etc.
Full object graph (functions, triggers, etc.)✅ (Enterprise)✅ (Pro / Enterprise )Atlas free limited to tables, indexes, FKs
Visualize schema as ERD❌✅ (Pro for private, free for public)Atlas also supports generating schema docs
Export schema to structured code folders❌✅Atlas supports exporting to organized folder structures
Generate ORM models from schema❌✅Atlas supports multiple ORMs (TypeORM, GORM, Sequelize, etc.)
Custom export formats using templates❌✅Atlas supports Go templates for custom outputs

Example: Compare without Snapshot​

The atlas schema diff command can also be used to compare two live databases without creating intermediate snapshot files.

atlas schema diff \
  --from "postgres://localhost:5432/staging?sslmode=disable" \
  --to "postgres://localhost:5432/production?sslmode=disable"

See comparing schemas for more details. Atlas supports comparing any source to any source.

Key Advantages over Flyway Snapshot​

  • Free: basic schema inspection is available in the free version.
  • Advanced filtering: use glob patterns to include/exclude resources by name or type.
  • Multiple output formats: choose between HCL, SQL, JSON, Mermaid, custom templating, etc.
  • Declarative model: output can be used directly in versioned workflows (atlas schema apply, atlas migrate diff).
  • Pro extensions: introspection for functions, triggers, procedures, and extensions.
  • Drift detection built-in: compare any two states deterministically.
  • CI/CD ready: easily automate snapshot generation and comparison in pipelines.
  • Export to code: generate organized folder structures with your schema definition.
  • ERD visualization: create interactive diagrams from your database schema.
  • ORM model generation: generate models for Go, Python, Java, and more from your schema.

Read More​