I'm using my Pixel 2 (non-XL) rather sporadically throughout the day, most of the time I just keep it in airplane mode with WiFi (and WiFi Call) enabled, except 1-2 hours outside, plus 30-60 minutes of active use per day.
Bought it shortly after release and so far I typically always had around 2 days of battery life per charge (varying +-6 hours or so).
Now and then the average battery estimate dropped, went back up, etc. as you'd expect from a "guesstimate".
However, over the last two or three weeks the estimate dropped lower and lower and this weekend I arrived at 24 hours while still dropping (even while fully idle and with full plane mode, mind you).
At first, I suspected maybe my router and/or repeater (providing both 2.4GHz and 5Ghz) are acting up or confusing the phone (like constantly swapping channels). However, this seems to not be the case as neither keeping full plane mode without WiFi nor turning off 5Ghz or the repeater changed anything.
In fact, I couldn't notice any significant difference between power consumption whether airplane mode is on or off (including WiFi, Bluetooth, and WiFi Call being on/off). No matter what I do, batterystats/Battery Historian (3.1 Docker Image) show an average drop of around 3-6% per hour while idle (face down). The drop is higher while used, obviously.
Apps wise I don't see any massive outlier either. Only CPU tasks running repeatedly throughout the day is qcom.glink-smem-native-xprt-rpm, typically running for about 5 seconds over 5 minutes. While the phone doesn't doze all the time, it's shown to regularly be in light or full doze. I don't see any regular long wake locks except com.google.android.gms once or twice over time.
However, one very odd thing that got my attention is the estimated/calculated Coulomb Charge shown in batterystats/Battery Historian (which also matches AccuBattery's estimate; assuming it does its standalone calculation).
Using one example bugreport I'm having at hand right now:
| Time since charging | Estimated Charge Left | Estimated Battery Percentage |
|---|---|---|
| +0 | 2615mAh | 100% |
| +1h | 2573mAh | 97% |
| +6h | 2386mAh | 77% |
| +10h | 2284mAh | 64% |
| +10h20m | 2178mAh | 59% (active use) |
| +12h | 2136mAh | 54% |
| +18h | 1987mAh | 35% |
| +22h | 1861mAh | 19% |
The last 20% would feel significantly longer than the previous percentages (as it should with the estimate being off), but it definitely doesn't last as long as 1800mAh should last in comparison (i.e. phone turning off, then only showing the static white battery screen). Two weeks ago the lower end of the scale sat at around 500mAh at 12%. Tried deep discharging twice so far, without the situation changing significantly; although that wasn't back to back).
Looking for "Coulomb Charge" unfortunately didn't really reveal any useful hints so far. Anyone experienced this before? What would be the next interesting thing to look at or try? Also noteworthy I'm not eager on just wiping the phone, having to once again setup authenticator apps etc.
