The landing dashboard is rebuilt as a command center, and none of
upstream's dashboard pieces remain: no tab strip, period dropdown, stat
cards or line-chart cards.
- One shell around every dashboard page, kept as you move between them:
navigation (Command center, then Trends, plus Cluster on a cluster),
the greeting band with live and node lamps, period keys, refresh,
auto refresh (off, 1, 5, 10 or 15 minutes; 5 by default) and a clock
with UTC and when the numbers last came in, then Needs attention.
- Needs attention replaces the status line: a tile per matter, critical
first, each opening where you'd deal with it; not drawn when all is well.
- Command center: dials for load, queue on schedule, mail received (to
inboxes against spam filtered), take-in time, nodes up and quota used;
quick stats with a strip of the period; live connection meters;
response times against fine/fair/slow marks; mail flow, weekly rhythm,
where mail is waiting and who uses the space.
- Trend pages (Network, Security, Delivery, Performance, Storage): each
gets its own dials and instruments over readouts and restyled charts.
- Cluster page when the server is one node of several: nodes up, oldest
heartbeat, how much of the period the links and data layer ran clean,
a live node roster, and error trends.
- A refresh keeps the last numbers on screen until new ones arrive, so
nothing blanks; instruments update in place.
jcoffey-dev