Right now you have graphs for CPU, Bandwidth, Disk, but that's only graphs and that's not quite enough.
I think you can provide some sort of "monitoring as a service" for droplets.
This can be done agent-less, by providing some standartized way (for graphs it can be just graphite-compatible input on your side). Or you can release optional agent.
Also there are some checks (TCP checks) that can be done server-side, and with them you could check for service availability (e.x. HTTP checks, SSH checks, mysql checks, etc - just to test if TCP connection can be established).
Droplet healthchecks are under considering for later in 2020. We are planning app/service uptime checks (url-based, ping) in the near term. Beyond that, we are exploring a more fully featured paid monitoring-as-a-service product, which would support custom metrics and visualizations.
Attachments Open full size
Also, what about a simple "powered on" check? I can't trigger an alert if my instance goes down. That seems like a pretty huge miss.
Attachments Open full size