Development Edition stops after a few hours?

[sbe-arg]

Describe the bug I have a weird behavior going on.

Every 2-3 days I have the need to edit documents. Each time I try to load any document I get the error for "Unable to load document" after 2-3 retries it starts working and works for a few hours trough the day without any issue and then the issue happens again.

Is there any sort of sleeping mechanism or warm-up required? But is the only think i can think of at the moment. Like if the Development server stops after X time and gets started only when a file is opened.

I don't mind If I it will waste a bit more resources but I'd prefer to have it running warm all the time.

To Reproduce Steps to reproduce the behavior: Open any file after 2 days of not using the editor.

Expected behavior Always open documents without taking several retries to start?

Screenshots

Client details:

• OS: Ubuntu 24.04 It happens on all clients.

Server details

Operating system: ubuntu 24.04

Web server:

Database:

PHP version:

Nextcloud version: 29 snap (but also occurs on 29 docker)

Version of the richdocuments app 8.4.4 - 24.04.5.2 ca2ed20

Version of Collabora Online 8.4.4 - 24.04.5.2 ca2ed20

Configuration of the richdocuments app

{
    "apps": {
        "richdocuments": {
            "disable_certificate_verification": "yes",
            "doc_format": "ooxml",
            "edit_groups": "",
            "enabled": "yes",
            "installed_version": "8.4.4",
            "types": "prevent_group_restriction",
            "use_groups": "",
            "wopi_allowlist": "172.16.0.0\/12,192.168.1.0\/24",
            "wopi_url": "https:\/\/cloud.somedomain.tld\/extra-apps\/richdocumentscode\/proxy.php?req="
        }
    }
}

Logs

#### Nextcloud log (data/nextcloud.log) ``` There are no logs capture by this behaviour. ```

[juliushaertl]

This sounds like an issue with https://github.com/CollaboraOnline/richdocumentscode so maybe you can file an issue over there. I generally would recommend a dedicated Collabora setup either using distribution packages or a docker container. While the "built-in CODE server" is an easy way to get it running, it is just spawning a process and there is not very much control over it if it gets killed for some reason.

[sbe-arg]

Will open one there.

But would be good to have more control or logs of the process with a verbose flag of sorts.