Kan (Kanban Board)

Date: 2026-06-24 Hostname: docker-host (LXC 100) IP address: 192.168.0.110

Self-hosted Trello alternative. Runs as a Docker Compose stack on LXC 100, managed via Komodo (GitOps).

Overview

Property	Value
Service	Kan (kanbn/kan)
Host	LXC 100 (docker-host)
Access	http://192.168.0.110:3006
Image	ghcr.io/kanbn/kan:latest
Database	PostgreSQL 15 (kan-db, internal only)
Network	arr_stack (external)

Stack layout

Three containers (compose/proxmox-lxc-100/kan/docker-compose.yml):

• kan - Next.js web app, published on host port 3006.
• kan-db - PostgreSQL 15, data bind-mounted to ${DOCKER_DATA}/kan/postgres. Not exposed to the host.
• kan-migrate - one-shot DB migration; runs to completion before kan starts.

Configuration

Secrets live in the Komodo Stack Environment (not git):

Variable	Purpose
KAN_POSTGRES_PASSWORD	Postgres password (shared by all three services)
KAN_AUTH_SECRET	better-auth secret, 32+ chars (openssl rand -base64 26 \| tr -dc 'a-zA-Z0-9' \| head -c 32)
KAN_BASE_URL	Access URL, must match how you reach the app (http://192.168.0.110:3006)

Email is disabled (NEXT_PUBLIC_DISABLE_EMAIL=true); enable SMTP vars from upstream if needed.

First run

The first registered user becomes the admin. Set NEXT_PUBLIC_DISABLE_SIGN_UP=true after creating your account if you want to lock signups.

Deployment

cd compose/proxmox-lxc-100/kan
docker compose up -d

Or via Komodo: Pull -> Deploy after committing changes.

Komodo shows "unhealthy"

Normal. kan-migrate is a one-shot container - it runs the DB migrations, then exits 0. Komodo counts any exited container as unhealthy even though this one is meant to exit. The app is fine as long as kan and kan-db are Up. Every Pull/Deploy re-runs (and re-creates) the migrate container, so the status returns; clear it with docker rm kan-migrate if the red status bothers you.