Alpha tests

[p1gp1g]

Please use the following template to provide efficiency feedback.

Several different factors can play a role in battery consumption. But we may get a good idea with many feedback. The objective is to compare FCM/background socket/unifiedpush with websocket strategy/unifiedpush with rest strategy.

Please not that the REST Strategy may reach rate limit on some public provider such as https://ntfy.sh.

* Your phone model:
* Your phone model's [don't kill my app](https://dontkillmyapp.com/) score:
* Operating System Version:
* MollySocket Version/commit:
* Molly (android) Version/commit:
* Molly (android) battery optimized (yes/no):
* UnifiedPush Distributor: (eg. ntfy/NextPush)
* Public server (yes/no): (eg. yes: https://ntfy.sh)
* Fetch Strategy (rest/websocket):
* Molly use time:
* Average battery consumption (molly / phone / duration): (eg. molly: 8%, phone: 70%, duration: 25h)
* Previous notification strategy (background socket/FCM):
* Missing notifications (no/one or two/a lot):
* Previous average battery consumption:
* Battery drain reduced/increase (how much):
* Anything else:

[p1gp1g]

• Your phone model: Pixel 6A
• Your phone model's don't kill my app score: 🏆
• Operating System Version: GrapheneOS - Android 13
• MollySocket Version/commit: 0.1.0
• Molly (andoid) Version/commit: v6.2.3-1-alpha_up1
• Molly (andoid) battery optimized (yes/no): yes
• UnifiedPush Distributor: NextPush
• Public server (yes/no): no
• Fetch Strategy (rest/websocket): REST
• Molly use time: 1h30
• Average battery consumption (molly / phone / duration): Molly: 3%, Phone: 80%, Duration: 39h
• Previous notification strategy (background socket/FCM): -
• Missing notifications (no/one or two/a lot): no
• Previous average battery consumption: -
• Battery drain reduced/increase (how much): -
• Anything else: -

Edit: I've seen a few notifications delayed later. I'm now trying without battery optimization.

[p1gp1g]

• Your phone model: Pixel 6A
• Your phone model's don't kill my app score: 🏆
• Operating System Version: GrapheneOS - Android 13
• MollySocket Version/commit: 0.1.0
• Molly (andoid) Version/commit: v6.3.6-1-alpha_up1
• Molly (andoid) battery optimized (yes/no): no
• UnifiedPush Distributor: NextPush
• Public server (yes/no): no
• Fetch Strategy (rest/websocket): REST
• Molly use time: 35min
• Average battery consumption (molly / phone / duration): Molly: 1%, Phone: 40%, Duration: 27h
• Missing notifications (no/one or two/a lot): no
• Anything else: No more delays

Note: I will add a wake_lock to the REST strategy and try with battery optimization again

[spazziale]

• Your phone model: POCO X3 PRO
• Your phone model's don't kill my app score: 🏆
• Operating System Version: DivestOS - Android 12
• MollySocket Version/commit: 0.2.1 ( Alpine Build )
• Molly (andoid) Version/commit: v6.9.2-1
• Molly (andoid) battery optimized (yes/no): no
• UnifiedPush Distributor: ntfy
• Public server (yes/no): no
• Fetch Strategy (rest/websocket): websocket
• Molly use time: 13h ( don't know the "real" usage)
• Average battery consumption (molly / phone / duration): Molly: 11%, Phone: ~100%, Duration: 13h
• Missing notifications (no/one or two/a lot): no

Note: I will test also with battery optimization again, as power usage seems too high

[p1gp1g]

@marc0x1

Thanks for your feedback !

Did you notice a better battery consumption compared to Signal/Molly with the background service (without UnifiedPush)?

Note: I will test also with battery optimization again, as power usage seems too high

You will get better battery result with the REST strategy.
Since you are using your own ntfy server, you do not risk rate limit.
(Battery optimization is good to get rid of unwanted services, but that's not the case here)

[spazziale]

Did you notice a better battery consumption compared to Signal/Molly with the background service (without UnifiedPush)?

Surely by a lot but since I have changed ROM I do not have old data about power consumption. Also since I use mollysocket in airgap mode I can use it only when I'm home.
For comparison, molly was always first on the list of power hungry apps.
Now I am switching to v0.2.2 and testing the REST strategy

[genofire]

@marc0x1 if your unifiedpush provider is accessible over the internet, the airgap mode should also work on the road (not only at home). So your molly can not communicate with mollysocket (directly to register), but mollysocket can send push notifications to your molly.

[fl0bo]

• Your phone model: Google Pixel 4A
• Your phone model's don't kill my app score: 🏆
• Operating System Version: LineageOS 20-20230803 / Android 13
• MollySocket Version/commit: 612be62
• Molly (android) Version/commit: v6.31.2-1+unifiedpush.rev8 (Molly-Insider-v6.31.2-1+unifiedpush.rev8-FOSS.apk
• Molly (android) battery optimized: no
• UnifiedPush Distributor: Conversations
• Public server: yes
• Fetch Strategy (rest/websocket): ?
• Molly use time: tbd
• Average battery consumption (molly / phone / duration): tbd
• Previous notification strategy (background socket/FCM): background socket
• Missing notifications (no/one or two/a lot): tbd
• Previous average battery consumption: "high" (was using non up-enabled version of Molly before)
• Battery drain reduced/increase (how much): seems like it is greatly reduced
• Anything else: leaving this for now after short use, will update later again

[fl0bo]

• Your phone model: Google Pixel 4A
• Your phone model's don't kill my app score: 🏆
• Operating System Version: LineageOS 20-20230803 / Android 13
• MollySocket Version/commit: 612be62
• Molly (android) Version/commit: v6.31.2-1+unifiedpush.rev8 (Molly-Insider-v6.31.2-1+unifiedpush.rev8-FOSS.apk
• Molly (android) battery optimized: no
• UnifiedPush Distributor: Conversations
• Public server: yes
• Fetch Strategy (rest/websocket): ?
• Molly use time: tbd
• Average battery consumption (molly / phone / duration): tbd
• Previous notification strategy (background socket/FCM): background socket
• Missing notifications (no/one or two/a lot): tbd
• Previous average battery consumption: "high" (was using non up-enabled version of Molly before)
• Battery drain reduced/increase (how much): seems like it is greatly reduced
• Anything else: leaving this for now after short use, will update later again

After using this setup for three weeks (updated to v6.35.3-1+unifiedpush.rev8 in the meantime) I can clearly state that battery drain is reduced from a about charge once a day to every two days, while I think I'm missing notifications from time to time.
I'm using Signal/Molly too rarely to allow for a better judgement.
For one case after a longer period of not using it, it felt like opening Molly triggered a notification for a message that was apparently sent an hour ago already.
@p1gp1g what would be the best way to analyze this? Can I run mollysocket in a kind of debug mode to match notifications with actual reception on the phone?
I have no idea if up.conversations.im is a good choice, or if using a self-hosted instance (ntfy?) would be better.

[p1gp1g]

Have you disabled battery optimization for Conversation too ? Any distributor should work the same.

You should not miss any notification, else : one service is being killed (that could be Conversation or Molly - this happen for FCM users too, but with vendors with aggressive battery "optimization") or Conversation server was down at a moment, which is probably not the case.

[fl0bo]

Have you disabled battery optimization for Conversation too ?

Yes, battery optimizations are disabled for Conversations.
Regarding "too" would it be fine to enable battery optimizations for Molly? (also disabled currently)

Status for Molly is total 14 Min.; 11 Min. in background. For Conversations total 1 Day, 10h, 18 Min; Day, 10h, 11 Min. in background.
41% left.

You should not miss any notification, else : one service is being killed (that could be Conversation or Molly - this happen for FCM users too, but with vendors with aggressive battery "optimization") or Conversation server was down at a moment, which is probably not the case.

OK