Kaneo - use a single application service instead of two (replicate Kaneo configuration)

[DimitriHautot]

Kaneo: fix services

Description

Recently, the Kaneo application code has been merged into a single service, so no distinct frontend + backend anymore, just one service.
Official documentation here.

Related Issues

• None.

Verification

Edited the.env file:

• added value for TS_AUTHKEY
• set my own tailnet in the KANEO_CLIENT_URL
• generated a random value for AUTH_SECRET

Then docker compose up

Checklist

• I have performed a self-review of my code and followed the templates structure.
• I have added verification that the stack works as expected.
• I have updated necessary documentation (e.g. frontpage README.md ).
• I have selected the correct label(s) for this PR.

Additional Context

• None.

[jackspiering]

Thanks for updating Kaneo to the bundled image, @DimitriHautot! I tested this branch locally: Tailscale, PostgreSQL, and Kaneo all become healthy. The web UI and /api/health return 200 over HTTPS through Tailscale Serve, and creating an account works. So the approach is good.

A few things before we can merge (see CONTRIBUTING.md and the service template):

1. Merge conflict. services/kaneo/compose.yaml now conflicts with main, which added health check comments and settings to the Kaneo services in the meantime. Please merge main again. When you resolve it, keep a single app service and add the comment # Healthcheck: defined by the image (wget against /api/health), so this file does not override it. The ghcr.io/usekaneo/kaneo image has its own HEALTHCHECK, so no healthcheck: block is needed.
2. Service key. CONTRIBUTING asks to keep the Compose service keys tailscale and application. Please rename kaneo: to application:. The container name app-${SERVICE} is already correct.
3. Image variable. Please use the template name IMAGE_URL instead of IMAGE_URL_SERVICE.
4. README. Please update services/kaneo/README.md for the single container: the service names in "Configuration Overview", and a note that KANEO_CLIENT_URL must be your https://kaneo.<tailnet>.ts.net URL. The upstream docs derive the API URL from it. Mention that the container has to be recreated after you change the value.
5. Optional cleanups:
   • Remove the unused top-level volumes: postgres_data:, because the database uses a bind mount.
   • Move the .env comments onto their own lines like in the current template.

Thanks again!

[DimitriHautot]

Hi @jackspiering ,

Thanks for your code review. I rebased my branch and pushed a new commit with the requested changes.
I hope it will be okay; otherwise I'll fix it again...

[jackspiering]

Thanks for the quick update, @DimitriHautot! All five points are addressed.

I tested the branch again. All three containers become healthy, and the web UI, /api/health, and sign-up work over HTTPS through Tailscale Serve.

Two small things are left:

1. Typo in services/kaneo/.env

SERVICE=kaneok should be SERVICE=kaneo.

With the typo, the node is named kaneok.<tailnet>.ts.net. That no longer matches KANEO_CLIENT_URL, and sign-up then fails with 403 Invalid origin.

2. Markdown lint in services/kaneo/README.md

Please add a blank line below the ### .env file and ### Container image headings. You can check it with:

rumdl check --config .markdownlint.yml services/kaneo/README.md

Optional: "TailScale" → "Tailscale" in the README.

After that, this is good to merge. Thanks again!

[DimitriHautot]

Hi @jackspiering ,

My bad, kaneok was my local test to not mess with my own instance of Kaneo. 😅

The linter has been applied to the README.md file, and the upper 'S' in TailScale has been lowered.

Thanks for your fast update !

[jackspiering]

@crypt0rr please have a second opinion, feel free to merge when approved.