> ## Documentation Index
> Fetch the complete documentation index at: https://docs.interaone.app/llms.txt
> Use this file to discover all available pages before exploring further.

# Update and upgrade

> Upgrade InteraOne safely with release review, coordinated backups, immutable images, health checks, and rollback planning.

Treat InteraOne as a coordinated application release: gateway, console, launcher, agent, and worker can share queue payloads, API contracts, database fields, and environment expectations.

<Steps>
  <Step title="Read and pin the release">
    Review the changelog and code changes, especially schema/index, queue, environment, OpenAPI, and deployment changes. Record the target tag or image digest for every service.
  </Step>

  <Step title="Back up coordinated state">
    Capture MongoDB, MinIO, Redis where recovery requires it, Qdrant or a plan to reindex, Caddy state, and the exact environment/configuration. Verify the backup before continuing.
  </Step>

  <Step title="Test in staging">
    Restore representative data, deploy the release, run build/API/widget/channel/AI/ingestion tests, and observe queue compatibility.
  </Step>

  <Step title="Pull and reconcile">
    ```bash theme={null}
    docker compose --env-file docker/.env -f docker/docker-compose.yml pull
    docker compose --env-file docker/.env -f docker/docker-compose.yml up -d
    ```
  </Step>

  <Step title="Verify and observe">
    Check container health, gateway health, login, tenant switching, widget assets, Socket.IO, one AI response, one background email, and queue failures. Monitor beyond the immediate deploy window.
  </Step>
</Steps>

## Rollback

A safe rollback requires previous immutable image tags and data that remains compatible. If a release performed an irreversible migration or vector rebuild, restoring old containers alone may not work. Document the data rollback path before deployment.

<Warning>
  Avoid `latest` in production. A restart can otherwise pull a different build from the one you tested, making incidents and rollback much harder to reproduce.
</Warning>

## After upgrading

* Rebuild/redeploy launcher assets when launcher or public URL settings change.
* Reindex knowledge after an embedding model/dimension change.
* Rotate secrets if the release addresses credential exposure.
* Update monitoring and runbooks for new queues, routes, or health checks.
* Update this documentation and `openapi.json` with the same release.