> For the complete documentation index, see [llms.txt](https://developer.harness.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://developer.harness.io/database-devops/3.0/use-db-devops/reference/glossary/tag.md).

# What is a Tag?

A **tag** in Harness Database DevOps is a marker that identifies the state of your database at a specific point in time. Tags represent versions, releases, or checkpoints and are essential for enabling targeted rollbacks and traceable deployments.

Tags are recorded in the `DATABASECHANGELOG` table, a Liquibase-managed audit table that tracks every [changeset](/database-devops/use-db-devops/reference/glossary/changeset.md) applied to your database. You can create tags using either the `tag` command (manual, applied at deployment time) or the `tagDatabase` Change Type (changelog-driven, version-controlled alongside your schema).

### Tag command <a href="#tag-command" id="tag-command"></a>

The `tag` command marks the current database state by writing a label to the most recent row in the `DATABASECHANGELOG` table. No changeset is created or executed; it is a metadata-only operation.

When you use the Harness Database DevOps apply step, tagging works in two phases:

1. The apply step checks whether the current database state is tagged before applying changes. If no tag exists at that point, it creates one automatically to guarantee a rollback target before any new changesets run.
2. After all changesets are applied, the apply step writes a post-deployment tag using the name you provided in the tag input field on the step configuration.

{% hint style="info" %}
The post-deployment tag is skipped when a deployment applies zero changesets. If no changes were deployed, no new tag is written. Go to [Tag Database Changeset](https://github.com/iKettles/harness-gitbook/tree/main/docs/database-devops/features/tag-database-changeset/README.md) to learn how to record a tag anchor even on no-op deployments.
{% endhint %}

![Tag input field on the Harness Database DevOps apply step" title="Click to view full size](/files/jUaCjx6mOFabOQtEbIMs)

### tagDatabase change type <a href="#tagdatabase-change-type" id="tagdatabase-change-type"></a>

The `tagDatabase` Change Type embeds a tag inside a changeset, making it part of your version-controlled changelog. When Liquibase executes this changeset, it inserts a new row into the `DATABASECHANGELOG` table containing the specified tag.

```yaml
databaseChangeLog:
  - changeSet:
      id: tagDatabase-example
      author: john-doe
      changes:
        - tagDatabase:
            tag: version_1.1
```

#### Behavior <a href="#behavior" id="behavior"></a>

* Inserts a new row into `DATABASECHANGELOG` with the specified tag value.
* Can be used as a target for `update-to-tag` (deploy only up to this tag) and `rollback` (revert all changes after this tag).
* Cannot be combined with other Change Types in the same changeset due to Liquibase XSD schema restrictions. Attempting to combine them causes a Liquibase validation error before any SQL is executed.

### When to use which <a href="#when-to-use-which" id="when-to-use-which"></a>

| Use case                               | Use `tag` | Use `tagDatabase` |
| -------------------------------------- | --------- | ----------------- |
| Manual tagging before rollback testing | ✅         | ❌                 |
| Changelog-driven version tracking      | ❌         | ✅                 |
| CI/CD-based release tagging            | ❌         | ✅                 |
| Creating lightweight checkpoints       | ✅         | ❌                 |

The `tag` command suits one-off, runtime checkpoints (for example, tagging before a hotfix). The `tagDatabase` Change Type suits repeatable, release-aligned tagging baked into your changelog.

### Best practices <a href="#best-practices" id="best-practices"></a>

* Tag after every application release to mark a known stable database state.
* Use semantic versioning for tag names: `v1`, `v1.1`, `v2`.
* Ensure tags reflect meaningful checkpoints so rollbacks target predictable states.

#### Sample tag flow <a href="#sample-tag-flow" id="sample-tag-flow"></a>

```
changeset1
changeset2
tag v1
changeset3
tag v1.1
changeset4
changeset5
tag v2
```

This structure allows rollbacks to specific versions based on real release points.

| **Tag**  | **Includes changes**   |
| -------- | ---------------------- |
| **v1**   | changeset1, changeset2 |
| **v1.1** | changeset3             |
| **v2**   | changeset4, changeset5 |

### Frequently asked questions <a href="#frequently-asked-questions" id="frequently-asked-questions"></a>

<details>

<summary>Can I roll back to a tag using Harness?</summary>

Yes. Use the rollback command with a tag name to revert all changesets applied after that tag. Harness supports rollback strategies via the pipeline rollback step.

</details>

<details>

<summary>Should I tag before or after a release?</summary>

Tag immediately after deploying a release. This creates a stable checkpoint with a name that matches the release, making rollback targets intuitive. The Harness apply step automatically creates a pre-deployment tag as well, so a rollback point always exists before new changes are applied.

</details>

<details>

<summary>Does tagging affect my database schema or data?</summary>

No. Tags are metadata entries in the DATABASECHANGELOG table only. They do not modify your database schema or data.

</details>

<details>

<summary>Can I preview which changesets will be rolled back by a tag?</summary>

Yes. Use the rollback SQL preview command or review Harness pipeline logs to see which changesets will be reverted before executing a rollback.

</details>

### Related concepts <a href="#related-concepts" id="related-concepts"></a>

Learn more about related concepts in Harness Database DevOps:

* [What is a Changeset?](/database-devops/use-db-devops/reference/glossary/changeset.md): Understand the unit of change that tags mark in the DATABASECHANGELOG table.
* [Using rollback tags with apply schema step](/database-devops/3.0/use-db-devops/deployment-pipeline-configuration/rollback-and-failure-strategies/using-rollback-tags.md): Step-by-step guide to configuring rollback targeting with tags.

### Next steps <a href="#next-steps" id="next-steps"></a>

* [Tag Database Changeset](https://github.com/iKettles/harness-gitbook/tree/main/docs/database-devops/features/tag-database-changeset/README.md): Configure the apply step to record a tag anchor even on no-op deployments.
* [Automatic and custom rollback](/database-devops/3.0/use-db-devops/deployment-pipeline-configuration/rollback-and-failure-strategies/automatic-and-custom-rollback.md): Set up rollback strategies in Harness Database DevOps pipelines.