Station health

A weather site is only as good as the hardware behind it, and a station that has quietly stopped reporting looks, on the public page, much like one that is working. Station health is the set of figures that tell the two apart. You always see them on your dashboard; you can also choose to publish them.

On your dashboard

The Overview page of a site shows a Station health card for every station, with a status badge:

The reporting interval is worked out from your archive rather than assumed, so a WeatherLink Basic station on a 15-minute upload is judged against 15 minutes and a gateway posting every minute against one minute.

Under the badge:

• Last reading: how long ago the newest reading arrived.
• 24h delivery: how many readings landed in the last 24 hours, and what share of the expected number that is. A day at 100% has no gaps; a day at 60% lost a good part of itself somewhere.
• Expected interval and the largest gap inside the last 24 hours.
• A low-battery flag when a transmitter has raised one.
• Whatever else the hardware reports about itself: console supply and battery voltage, solar panel voltage, network signal strength, station and link uptime, and on a relayed Davis console the network error, CRC error, resync and missed-packet counters.
• The sensor groups detected on the station, so you can see at a glance that the soil probes or the lightning detector are being recognised.

View sensor details takes you to the Connection page, which lists every discovered sensor.

Which of those figures exist depends on the hardware. A Davis console can report supply voltage and uptime but not a battery percentage; an Ecowitt gateway the reverse. Anything a station does not report is left out rather than shown as a dash.

On your public site

The same figures can appear as a Station health tab for visitors. Turn on Station health under Panels → Tabs & sections. It is off by default, because telling the world how your hardware is doing is a deliberate choice, and switching it off never hides anything from your own dashboard.

The public tab shows a status (Online, Late or Offline), the last reading and reporting interval, data completeness with the count of readings in the last 24 hours, the largest gap, and each battery, signal, voltage and uptime figure your station supplies. Every tile hides itself when its value is missing, for the reason above: a fixed grid would show most stations a wall of dashes and imply faulty hardware where there is only quiet hardware.

Being told when something is wrong

You do not have to keep looking. Three station health alerts watch these figures for you:

• Data older than a duration you choose: the alert for a station that has gone quiet, and worth setting on every site.
• Battery below a level, for hardware that reports one.
• Reception below a percentage, for Davis consoles that report it.

These three are the only alerts still evaluated when a station has stopped reporting, since a temperature threshold checked against an hour-old reading would not be a measurement. Set While it lasts on the data-age alert to be reminded every few hours until the station is back.

Next steps

• Alerts & warnings
• Connect other hardware: what the ingest card shows about uploads
• Troubleshooting