Upgrading Appaloft

Goal

Safely upgrade a self-hosted Appaloft instance to a new version, and confirm every capability works correctly afterward.

When to use this task

You need new features or a security fix.
The Web console’s Instance page shows a new version is available.

Prerequisites

You’ve already completed a pre-upgrade backup — see Database Status And Migrations.

Inputs and defaults

Upgrade order: back up state first, then upgrade the binary or image, and finally confirm the Web console, /docs/*, database state, and provider/plugin status are all normal.

CLI steps

Checking for updates

appaloft upgrade check

The Web console’s Instance page also shows the current version, latest version, target version, and a directly copyable SSH update command.

Applying the update

A standard self-hosted install can just re-run the installer directly — it reuses /opt/appaloft/.env and the existing data volume:

curl -fsSL https://appaloft.com/install.sh | sudo sh

Re-running also aligns the Traefik proxy Appaloft manages to the current release’s reviewed default image. If APPALOFT_TRAEFIK_IMAGE is explicitly set, or an external proxy is being reused, that image remains the operator’s own responsibility to review and upgrade.

You can also target a specific version:

appaloft upgrade apply --version 0.2.1 --confirm

Expected output and status

The Web console’s “Update now” button is only enabled when the host-side process explicitly sets APPALOFT_INSTANCE_UPGRADE_APPLY_ENABLED=1 — a default containerized deployment doesn’t allow the Web process to directly modify the host machine’s Docker install; this is an intentional safety boundary.

Verification

Refer to the verification checklist in Database Status And Migrations: check the project list, recent deployments, environment variable snapshots, log entrypoints, and access URL status.

Rollback / recovery

Before rolling back, confirm whether the database migration is reversible and whether the old version can read the current state — some schema migrations are one-way, and you must restore the database from the pre-upgrade backup before rolling back the version, rather than simply redeploying the old binary.