Same here. I’m the only user of my services, so if I try visiting the website and it’s down, that’s how I know it’s down.

I prefer phrasing it differently, though. “With my current uptime monitoring strategy, all endpoints serve as an on-demand healthcheck endpoint.”

One legitimate thing I do, though, is have a systemd service that starts each docker compose file. If a container crashes, systemd will notice (I think it keeps an eye on the PIDs automatically) and restart them.

Thanks, this is a great explanation. I’ll try doing &> /dev/null & tomorrow. I’d like to find the simplest version of this for recommending to other people.

Thanks for the suggestion; I tried both of these things and they do hide the output, but it doesn’t make the post-receive script actually exit early. So git push still takes a while to finish.

I think this might have worked, except the ssh key I’m using for git has a password on it. So it does return to the terminal right away, but the git push command in the background is stuck waiting for me to type my password.

My JSON export from wallabag is 46 megabytes. That’s for 2,465 articles.

I love how active the development on Linkwarden is. I still have all of my stuff in wallabag, but Linkwarden is tempting. I gave the hosted trial a try a few weeks ago, but my wallabag export was too big to import. Maybe I’ll try selfhosting it and manually increasing the max upload size this time.