A dashboard of fourteen panels, replaced by six, chosen by reading a quarter of incident messages.
the questions actually asked during incidents:
"is it everyone or one customer?" 6 times
"when did it start?" 6
"did we deploy anything?" 5
"is the database or the app?" 4
"is the queue moving?" 3
"is the supplier up?" 3
six panels, in that order, on one screen.
the fourteen panels it replaced answered none of these
directly. two of them answered the last one indirectly.
A dashboard built from available metrics answers what is measurable and a dashboard built from questions answers what is asked, and those overlap less than expected. Deploy markers on the time axis were the single highest-value addition and are not a metric at all.