Skip to main content
CapRover apps import over its HTTP API, read-only. The main difference from the other sources is that CapRover builds images on its own host and does not expose them, so every app needs a source here: a git repository or a registry image. Read the migration overview first for the flow and the downtime table.

​
Prerequisites

  • Levelrail installed on the new server and the CLI signed in (levelrail-cli auth login).
  • The CapRover login password. It is exchanged for a session token with POST /api/v2/login (the only non-GET request the importer makes).
  • The CapRover URL reachable from the Levelrail control plane. For a private address, set APP_IMPORT_ALLOW_PRIVATE_NETWORKS=true and pass --allow-private.
  • Each app’s git repository, or a registry you can push its image to.

​
What imports automatically

​
What needs a manual step

  • A source for every app. Connect the repository (with credentials for a private one) and run a build, or push an image to a registry and set it as the source.
  • Volume contents, by rsync commands you run.
  • Raw host port mappings are not imported. Publish a port in the app settings if it must be reachable.
  • Health checks. CapRover has none to import. Add a readiness path here.

​
What cannot be migrated

  • Images built on CapRover: there is no recoverable image through its API. Rebuild from source.
  • Databases as a feature. CapRover runs databases as ordinary apps or one-click apps, so nothing maps to a managed database. Create a managed database here and copy the data.
  • Certificates, DNS records, deploy history, logs and metrics.

​
Steps

1

Dry run

2

Import

Apps are created without a running image. The report names the repository and branch where one is known. For verdicts and one-click fixes, use levelrail-cli migrate plan --platform caprover --url https://captain.example.com, then add --apply --only ID.
3

Give each app a source and build

Connect the git repository or set the registry image on the app, then deploy it. See Importing apps.
4

Move database data

Create a managed database of the same engine, with the same or a newer major version. Then copy:
The CapRover database container must publish a port the Levelrail node can reach. Publish it temporarily and restrict it by firewall. Point the app at the managed database and set its connection variable.
5

Copy volumes, check, switch

​
Databases and the downtime window

The database copy is a full dump and restore with row counts compared, repeatable, with the source read-only. There is no incremental sync. Pause the old app, run the final copy, start the app here, switch DNS. The window is the final copy time plus a restart. Apps without a database and without volumes can move with no outage you would notice. See the downtime table.

​
Domains and TLS

CapRover runs nginx on ports 80 and 443. Point DNS at a new server, or install Levelrail in coexist mode. Lower the TTL to 300 seconds a day ahead and let levelrail-cli migrate cutover tell you when it is low enough. Certificates are issued here on first use once DNS points here. See Domains and ingress.

​
Common problems and fixes

​
Choose to stay if

CapRover’s one-click apps and simple Swarm setup already cover what you run, and you do not want to re-point apps at a git source. Levelrail has no recorded run against a live CapRover. See CapRover alternative and the comparison.
Last modified on October 20, 2026