Alert noise, cut down to what matters
This is a sample entry. It shows the layout and is not a real write-up.
The idea
Group alerts by the thing that is actually broken, not by the check that noticed it.
1
Group by host
Better than nothing, but one bad switch still paged once per server behind it.
2
Group by dependency
In progress. Needs a dependency map I trust, which turns out to be its own project.