No logs from random containers after updating to 0.19.0

[hugalafutro]

Describe the bug I updated lazydocker on several pc/vm's and raspberry's and noticed when scrolling down several containers do not output any logs. docker compose logs -f or dozzle correctly display these. (Maybe related, but the installer script never worked properly for me, even when ran as root it ends up in home directory, so I usually run the script, then sudo mv ~/.local/bin/lazydocker /usr/local/bin)

To Reproduce Steps to reproduce the behavior:

1. open lazydocker 0.19.0 on host with multiple containers running
2. start scrolling down the list of containers
3. look in the log window, some of the containers will show just blank window with no logs whatsoever

Expected behaviour Logs should display for all containers

Screenshots I can make a video if necessary, but I think it's pretty self-explanatory - you should see logs, there's nothing.

Desktop (please complete the following information):

• OS: Ubuntu 22.04.1 LTS
• Lazydocker Version 0.19.0

Additional context I tried to find out some connection between the containers that don't work, but came up empty. It's random containers each restart. I have tried running lazydocker as root, but no change. This is not a slow cpu issue either I can stay on the container for 5 minutes the logs will never show.

[StefanoOcchetti]

Same happens to me, running LazyDocker with docker:

docker run --rm -it -v //var/run/docker.sock:/var/run/docker.sock lazyteam/lazydocker

[hugalafutro]

I subjectively feel it got even worse in 0.20.0, now it's about half or more of the containers displaying no logs in lazydocker.

[jesseduffield]

Right you are @hugalafutro: as of v0.20.0 we only show the last hour of logs for a given container (see commit here: https://github.com/jesseduffield/lazydocker/commit/0b7017150c7cf0534c71203a53c2ff0f608e459c).

This is configurable: you can set:

logs:
  since: ''

in your config to un-set the 'since' value that gets passed to docker logs.

The motivation behind that was that if a container has millions of lines of logs from being up for a long time, it was very painful to wait for all those logs to be produced so that you could get to the end.

But if that default is causing issues I wanna find a better solution. One approach would be to have a default value for the 'tail' arg of something like 1000 so that it only loads those lines and then it continues from there. But even 1000 takes a while to load for some containers I'm testing that on.

Something we could do is a similar approach to k9s which shows you the args within the app and provides keybindings for you to change them.

[jesseduffield]

I've got a WIP PR up for making this stuff visible and configurable from within the app https://github.com/jesseduffield/lazydocker/pull/406

[hugalafutro]

Thanks, I was not aware of this, it is actually an useful feature, I have containers that keep logs for months and it took like 10mins for it to scroll to the end on raspberry pi.

I was just taken aback by the blank screen that to me looked like bug, perhaps a line of text in such a case explaining that no logs have been produced in last hour could be useful, although in that case I would possibly still be interested in seeing the tail end of the last log instead of nothing.

I wouldn't say the defaults cause issues, it's just not readily apparent they changed and why in some cases there's no log (I presume large portion of user base doesn't read changelogs, I too am guilty of this at times - at other times I try, but it's hard when the log is long-ish and 90% of issues fixed/features introduced usually do not apply to my use case).

I'll be happy with whatever approach you'll choose as now that I know no logs means all's well it won't bother me that much.