Logs and system monitor

The Logs page shows the firmware's own log, kept in a ring that survives a restart, so the lines from before a crash are still there afterwards. The System Monitor shows what the processor and the memory are doing. Both are the first stop when something behaves oddly, and what support will ask for.

Logs

Logs. Errors are highlighted in red, warnings in amber, the `---- boot ----` separators in grey.

The ring holds the last 16 KB of log lines in memory that a warm restart (a reboot, a crash, a watchdog) does not clear; only a power cut does. Each start writes a ---- boot ---- line, so the ring reads as a short history of the last few starts. The page refreshes every 2.5 seconds.

Control	What it does
The tag field and level	Set the level of one component (wifi_manager, lte_upstream_pppos, gps, or * for all) to error, warn, info, debug or verbose until the next restart. The level stored for every start is All Settings › Logging › Level.
auto-scroll	Keep the newest line in view.
Download	Save the ring as a text file: what to attach to a support request, together with the About page.
Clear	Empty the ring.
The sink chips	console (the serial console) and ring: click one to turn it off or on until the next restart. The dropped count says how many lines were lost under a burst.

On the console debug 1 turns the live log output on and debug 0 off.

System Monitor

System Monitor. The CPU column is measured between two refreshes, so it reads after the second one.

Part	What it shows
The tiles	CPU load over the last 2 seconds, the number of tasks, free internal memory (with the lowest ever), the die temperature.
Internal RAM and PSRAM	The two memory pools: the small internal one holds the radios and the drivers, the large PSRAM everything else. Largest free block below Free now shows fragmentation.
Tasks	Every task with its state, core, priority, stack headroom (bytes never used: a small figure means the task is close to its limit) and CPU share.

Nothing here needs attention on a healthy dongle: a CPU load of a few percent, internal free memory above 100 KB, a die temperature in the 40s to 60s in a car. A rising Unexpected resets count on the Status page, not these figures, is the sign of a problem: see Crash report.