I run about 15 services in my homelab and I keep running into the same problem - something breaks visually and I dont notice for days because my uptime monitor says everything is fine.

Latest example: Nextcloud updated and the login page started showing a PHP warning at the top. Functionally everything worked, users could still log in, but it looked broken and unprofessional. Took me a week to notice because I rarely actually open the login page myself.

Got me thinking about automated screenshots. Set up a cron job that takes screenshots of all my service login/dashboard pages daily, then I can just scroll through them to spot issues.

Anyone doing something like this? I know there are commercial tools for this but wondering if theres a good selfhosted option or if people just script it with Puppeteer.

My main concern is resource usage - running headless Chrome for 15 URLs daily on a machine thats already doing a lot of other stuff.