As for the graphs, you will probably want to be aware that there are three types that are supported, two of which are historical (the "classic" and "classic - simple" ones, which weren't removed not to break anything) and the remaining type being more customizable and a bit more dynamic. Here's an example of all three types, the bottom one being the interactive kind: https://i.imgur.com/avhPoCO.png
Here's a gallery of the customization options, though most of them weren't relevant for me, because i didn't necessarily want to customize the graphs much in my dashboard, given that there are tens of them in there (i just wanted something to show some key metrics for all of my servers in a single page): https://imgur.com/a/9Ervntg
Oh, also the network map (with information about host problems) and web monitoring functionality is pretty nice, since i can also use Zabbix for checking how responsive all of the sites that are actually hosted on the server are and whether anything has gone down.
Overall, it's a pretty low maintenance piece of software with sensible defaults in regards to what it monitors and what causes alerts to be created. I only wish that it would be easier to customize it further (e.g. aggregating data at slightly longer intervals to store less data about my homelab, though their mass update functionality is pretty okay; or automatically creating graphs for new nodes), or that it would allow for e-mail/other alerts for web monitoring out of the box, as a default. Also, when configuring maintenance periods (because i don't want alerts about my servers going down when i've scheduled them to do that), the time zones etc. can indeed be a bit confusing and the UI is generally a tad clunky.
Also, it runs on pretty boring software (MariaDB/MySQL for the data) and the system requirements aren't too demanding, so i don't need a separate node, merely a few containers with about 512 MB of memory and 1 CPU core for each.