Netdata, looking at dashboards when something breaks :)

Maybe the people asking for your projects and the ones slamming them for you using AI in any capacity are different people.

I for one think AI is a tool and if you’re competent or it’s for one-off tasks, you can use it. It’s the agentic vibecoding by complete non-IT people (or techbros, or absolute beginners) that grinds my gears.

😱😬🤣😐🙃🙈💩

Maayyyyybe because he is?

I mean, if far-left radicals call someone a fascist, that’s debatable. But everyone left of center? Maaaaaybe they’re onto something here?

Did you mean ricing?

You mean 5W?

What do you use as reverse proxy? I have been trying to get a safe reverse proxy going via traefik in --user mode but it’s been rough 😄

Edit: checked your community, Caddy… Never got warm with that one…

but without the command

Yeah, but that’s an easy one to parse, if you really need it.

I’m pretty sure if you use OnCalendar it will still be there after a reboot. Let me check.

EDIT: oh damn!! It doesn’t persist on reboots. Sorry for this oversight.

If you inspect the generated timer you see the same format in the OnCalendar line.

What do you mean by persisted? The service and timer will both be there as long as the timer doesn’t end. Which it won’t in my example.

If you want to copy it for later reference, you can simply copy the files.

windows, yes: https://learn.microsoft.com/windows/wsl/systemd

for Mac you should probably use macOS’s launchd services: https://www.launchd.info/

single line

here’s a single line for you which creates a one-shot service+timer in one go, suspending the system in 4 hours:

systemd-run --on-active=4h --unit=suspend-after-4h /usr/bin/systemctl suspend

If you want it to run repeatedly, simply replace --on-active with --on-calendar="*-*-* 04:00:00" for example.

But why does it need a cross platform package when systemd is available everywhere? What UX is better than the already existing cockpit?

Yes but… Systemd timers are INI-defined tasks and services with unified history and logs, the alerts can be easily added with any alerting system.

Looks good, but kind of superfluous if you just use systemd timers, lol

Systemd timers are the modern way of doing it which gives you everything you coded this tool for, logs and whatnot, in the journal.

And you can run them in a web UI like cockpit aswell.

This might be interesting in a business setting but lacks ACLs and user management in general.

It’s probably also quite easy to hack the system with it if you can inject some malicious scripts.

Get better cooling. Wife and kids deserve an AC in these inhumanely hot times.

Downclock the system to have it run cooler for now.

Get a properly monitored backup system which either sends you and wife an email automatically sorted into a separate mail folder via IMAP sieve rules, or has a dashboard with last backup time and diff size which has triggers if the backup is older than 3 days or diff size is larger than 10% of total storage.

I recommend the LGM stack (Loki, Grafana, Mimir).

Document everything and make it as easy to retrieve the plain files as possible in case you die.

What’s the failure mode? Higher operating temperatures?

download the entire netflix library forever.

I’m pretty sure this is larger than 14TB?!

Honestly if you use Claude just tell it to set it up for you.

Not the point of a homelab and additionally very dangerous as Claude tends to like insecure configurations.

If you use TLS like you should, your domains will be on the internet in the certificate transparency log. Yes, you should use a wildcard cert if you want this security by obscurity, but it’s still security by obscurity.

My brain