ntfy.DOMAIN throws 404

[felixx9]

**Description
Call ntfy.DOMAIN in Browser throws 404:
{"code":40401,"http":404,"error":"page not found"}

To Reproduce
open https://ntfy.DOMAIN in browser.

ntfy_enabled: true
matrix_server_fqn_ntfy: "ntfy.{{ matrix_domain }}"
ntfy_configuration_extension_yaml: |
  log_level: DEBUG

Expected behavior
Until somewhen earlier (before introducing Traefik?) a page like https://ntfy.sh/app was served under ntfy.DOMAIN

[spantaleev]

This webapp is intentionally disabled, see: https://gitlab.com/etke.cc/roles/ntfy/-/blob/a53d7c80fb269d23b2675b9bff9e857a05525d66/defaults/main.yml#L101-108

Related to: binwiederhier/ntfy#256

Because you're not serving ntfy at a subpath but at / on its own domain, you may be able to get the webapp to work by adding ntfy_web_root: app to your configuration. People who are serving ntfy at a subpath are not so lucky, until that ntfy issue gets fixed.

The question is.. do you use or need this webapp? Serving a weird scary JSON message is not ideal though

[Zocker1999NET]

I found the web app very useful and was disappointed when it stopped working. Until finding this issue and the "explanation" I considered this being a bug. Thankfully I found this issue after a further playbook upgrade didn't fixed it automatically.

I find the web app useful as I use ntfy for "pager-like" notifications I not only want to receive on the phone but also on the desktop. And as I already self-host ntfy, I prefer to use "my own" web app version instead of the "official hosted" one so I do not depend on it being up.

I can understand the reasoning behind changing the default as most users are expected to use it as UnifiedPush service for Matrix or, if it being used for "pager"-like notifications, they may be bridged to Matrix using Maubot or so. So letting the web app enabled by default may is something most users won't expect.
But I still think this change should have been at least documented in the CHANGELOG of this repo because it came out of nowhere. It may still even be added to the CHANGELOG & to the ntfy documentation page so others know how they can (re-)enable the web app again.

[spantaleev]

This webapp was never advertised and I doubt that many people discovered and found a use for the webapp (like you did), so it wasn't mentioned in the changelog.

In the whole month after its removal, this issue (with 2 people mentioning it) is the only complaint I've seen, so it doesn't seem like a big problem.

[Zocker1999NET]

This webapp was never advertised

I understand what you mean, the web app was not explicitly advertised from this playbook. But the typical phrase at top of the service documentation is ambiguous, as it just refers to the ntfy homepage.

The playbook can install and configure the ntfy push notifications server for you.

And the ntfy homepage advertises the web app clearly. together with the top phrase "It's also open source if you want to run your own.".

and I doubt that many people discovered and found a use for the webapp (like you did), so it wasn't mentioned in the changelog.

In the whole month after its removal, this issue (with 2 people mentioning it) is the only complaint I've seen, so it doesn't seem like a big problem.

I can understand why this wasn't mentioned in the CHANGELOG until now, but I see no reason to not mention it now in the documentation and also the CHANGELOG. I see the chances are high that after one month, many playbook users haven't upgraded as well and so didn't ran in this issue for now. It may be even that users have found your comment on how to restore the web app and applied that for them but haven't written anything here (as just writing "me too" is not that accepted on GitHub for good reasons).

I will set up a PR for an addition to the ntfy documentation and a CHANGELOG entry.