Guide
Availability checks, server-resource history, notification delivery and a shared dashboard serve different jobs. Build only the layers you need and monitor from outside the server.
Configure an availability check for an endpoint you care about. Stop that service briefly, verify the failure is recorded, restart it, and verify recovery. Then connect a notification channel and repeat the exercise. A configured channel is not evidence that the phone actually received anything.
Add Beszel when historical CPU, memory, disk and network information will help explain failures. Install and restrict agents separately. DNS logs and Home Assistant sensors can add context, but do not replace endpoint checks or resource collection.
A monitoring service on the same machine cannot reliably report that machine losing power. Put at least one check on separate hardware or another server and consider shared power, network and notification dependencies. This guide lists compatible roles; it does not ship a prewired monitoring stack or claim tested delivery between every pair.
Sets
The sum of the minimums the recipes below declare. It is a floor, not a recommendation: leave room for the operating system, backups and growth.
Why it is here: Choose this when you want to manage checks through a web interface and publish a status page. Notification integrations still need separate configuration and a real delivery test.
Why it is here: Choose this for checks described in configuration files and a compact status page. Back up the configuration and state; it is an alternative to, not a required companion for, Uptime Kuma.
Why it is here: Adds resource history through separately deployed agents. The hub backup retains databases and SSH identity. The recipe test used an isolated agent, not a complete physical-host monitoring deployment.
Why it is here: A delivery destination after the producer is connected. Protect topic permissions and tokens. Known attachment URLs are bearer links; the recipe does not establish private attachment downloads or background mobile push.
Why it is here: Organizes links to the other tools and optional integration widgets. It has a shared password in the recipe, but does not replace an availability monitor or automatically configure integrations.