A changed WAN interface reaches the dashboard, and quieter CPU and ping alerts
Fixed
- Changing a router's Default WAN Interface now takes effect at once. The setting was saved,
but an open dashboard kept the old interface for its WAN status and default graph until
MikroDash restarted. Browsers already open move to the new interface; one that picked another
interface by hand keeps it. Reported in
#158, thanks
@elpanda88. - Switching routers puts the graph on the new router's WAN. It could land on that router's
first running interface instead, or show one interface in the dropdown and another on the chart. - A recovery is sent only to channels that were sent the alert. A channel set to alert at 90%
CPU was told "CPU back to normal" about short spikes under its threshold that it never heard
about. A breach held back by a channel's cooldown gets no recovery either.
Changed
- High CPU and Ping Loss alerts must persist before they fire. A reading has to stay high for
the router's Offline threshold (Settings, Devices, the router's "Offline threshold, seconds",
30 s by default), and back to normal as long before it recovers, so brief spikes no longer alert
or ring the bell. Set it to 0 to react at once, as before. - A channel hears about CPU or ping loss that climbs past its threshold mid-alert. An alert
that opened at 72% and then held at 95% now reaches a channel set to 90%.