Fix false recharge-cycle detection from a single noisy battery reading

A report was flagged as "the battery got recharged" (resetting
battery_history and stats_recharge_cycles, and re-arming the low-battery
alert) whenever it came in >= RECHARGE_JUMP_PCT above the single
immediately-previous report. That's exactly what a real recharge looks
like, but it's also exactly what a normal reading looks like right
after one noisy low report: e.g. 60, 59, 58, then a stray 53, then back
to a perfectly normal 58 -- 58 >= 53+5 falsely read as a recharge.

Now compared against the max of the last RECHARGE_LOOKBACK (3) reports
instead of just the one before it, so a lone stray reading doesn't get
to set the bar a normal reading then trips. A real recharge still needs
to clear all of them, so genuine recharges are still caught immediately
(verified: 18% -> 90% still triggers, history still resets).

Paired with the firmware-side battery.c change (trimmed-mean ADC
sampling) that reduces how often a stray reading like the 53 above
happens in the first place.

server/app/routers/common.py

@@ -21,7 +21,16 @@ logger = logging.getLogger(__name__)
 # Battery-history / estimate tuning (see /frame/battery and battery_estimate_s).
 BATTERY_HISTORY_MAX = 500  # ~20 days at hourly reports
 BATTERY_LOG_MAX = 20000  # ~2 years at hourly reports -- cap on the battery_log table per frame
-RECHARGE_JUMP_PCT = 5  # a report this much above the previous one = battery was recharged
+RECHARGE_JUMP_PCT = 5  # a report this much above the recent baseline = battery was recharged
+# How many of the most recent reports make up that baseline. A lone noisy
+# reading (ADC/regulator glitch -- see firmware/main/battery.c) can still
+# dip or spike a single report; comparing against just the one immediately
+# previous report meant that a normal reading right after a noisy dip
+# looked like a 5%+ jump and falsely registered as a recharge. Comparing
+# against the max of the last few reports instead means an actual recharge
+# still needs to clear all of them, while a single stray low one doesn't
+# get to set the bar.
+RECHARGE_LOOKBACK = 3
 MIN_ESTIMATE_SPAN_S = 2 * 3600  # need at least this much observed time...
 MIN_ESTIMATE_DROP_PCT = 2  # ...and this much observed drop before estimating