Backup & Restore CapRover

What Is Backed Up

A CapRover backup copies /captain/data and adds metadata about cluster nodes. This includes CapRover configuration and state, such as app definitions, domain settings, and certificates stored there.

• It does not automatically include Docker volumes, host bind mounts outside that directory, database contents stored in those locations, uploaded files there, or locally built Docker images.

• Back up each database using its own consistent export or snapshot process, and copy the persistent directories you actually use.

• Test that both the CapRover archive and the application data can be restored.

• An exception to the image rule is a self hosted registry: its stored image data is under /captain/data/registry, so that directory is copied and may greatly increase archive size.

• With an external registry, keep its credentials and image retention policy available for restoration.

• Without accessible images, redeploy apps after restoring their definitions.

See Create a Backup, Restore a Server, and Persistent Directories.

Create a Backup

1. In the CapRover dashboard, go to Settings and select Create Backup.
2. Download the resulting .tar file, verify it is nonempty, and store it outside the server with access controls.
3. Back up application databases and persistent directories separately; the CapRover archive has limited scope.
4. Periodically restore both kinds of backup to a test server and verify the applications.

• The archive is generated on the server, then offered as a temporary download.
• Keep your own copy: the generated download is temporary.
• If a self hosted registry is enabled, allow space for its image data in the archive.

Automate Backups

• The dashboard's Create Backup action has a corresponding authenticated API endpoint.
• The following example needs curl and jq.
• Set CAPROVER_URL to your HTTPS dashboard origin and provide CAPROVER_PASSWORD securely through your scheduler's secret store.

set -euo pipefail
umask 077
api_token=$(curl --fail --silent --show-error "$CAPROVER_URL/api/v2/login" \
  -H 'x-namespace: captain' -H 'content-type: application/json' \
  --data "$(jq -n --arg password "$CAPROVER_PASSWORD" '{password:$password}')" | jq -er '.data.token')
download_token=$(curl --fail --silent --show-error "$CAPROVER_URL/api/v2/user/system/createbackup" \
  -H "x-captain-auth: $api_token" -H 'x-namespace: captain' \
  --data '{"postDownloadFileName":"backup.tar"}' | jq -er '.data.downloadToken')
curl --fail --silent --show-error -G "$CAPROVER_URL/api/v2/downloads/" \
  --data-urlencode 'namespace=captain' --data-urlencode "downloadToken=$download_token" \
  -o backup.tar
test -s backup.tar

• Move the file to dated off-server storage, encrypt it as appropriate, and set retention and failure alerts.
• The download token is temporary; finish the download in the same run.
• Schedule separate database and volume backups and regularly test restoration.

Backup Contents

The backup implementation archives /captain/data and adds meta/backup.json with the CapRover salt and cluster node metadata.

Included	Excluded unless stored inside /captain/data
Saved CapRover configuration, app definitions, certificates, and self hosted registry data under /captain/data/registry.	Docker volumes and host bind mounts elsewhere, database files there, and local Docker image store.

• The registry exception can make archives large.
• The restore code looks for /captain/backup.tar at initial installation and expands it into /captain/restoring.
• See What Is Backed Up and Restore a Server for the full recovery sequence.