Support for non Docker log files?

[laneyweb]

Firstly, thank you so much for Dozzle. It makes reading/checking Docker logs so much easier.
I would like this functionality for remote log files over SSH for non Docker systems.

Is this possible in Dozzle? If not, can anyone recommend an alternative that allows me the same sort of log access but using remote SSH access to view specified log files?

Thanks again for Dozzle!

[amir20]

https://dozzle.dev/guide/log-files-on-disk

[NatoBoram]

This one is so strange. It doesn't feel right to create separate Docker containers, one per log file we want to view.

Out of the box, my homelab already has a use case for this.

• Some Docker containers have multiple logfiles, like access logs vs error logs.
• What if I want Caddy to output a different stream per subdomain?
• There are apps that are only compatible with logfiles so I have to use that, but I also want to view them.

And more recently, I started using the agent on a different machine. It's a Digital Ocean droplet running a WooCommerce website. I'm able to read the logs of the recently added Docker services put on that machine, but that machine used to be very under-powered and I was only able to start adding Docker and stuff on it since a recent hardware upgrade. So now I have new fun problems:

• That machine already has an ecosystem of logfiles that I want to monitor in addition to the new Docker-based services
• There's a lot of logfiles that can be generated by WordPress/WooCommerce
• Apache also has its own ecosystem of logs
• I also want to obtain system logs

At this point, one might say that I need a proper log aggregator, but they're always so much more complicated and split into 12 different microservices. Dozzle already has the agent+auth+docker+hub+ui all-in-one and the only missing link is the ability to read logfiles.

I really think this is a legitimate use case for Dozzle.

[amir20]

I also think it's a legitimate use case. But Dozzle can't do every thing and it heavily relies on the Docker API. Implementing all the functionality that would be required to read, stream and fetching historical log would be too much work.

I personally prefer using Docker when possible and send everything is to stdout/stderr. Writing to a file is an anti-pattern.