High battery usage #531

@Ginner commented on GitHub (Dec 6, 2023):

Might be related to my issue here: #1140.
Do you know if you're changing networks?

@Ginner commented on GitHub (Dec 6, 2023): Might be related to my issue here: #1140. Do you know if you're changing networks?

saavagebueno commented

@Pounii25 commented on GitHub (Dec 6, 2023):

Might be related to my issue here: #1140. Do you know if you're changing networks?

Do you mean If I change the provider 5g antenna during the day?

@Pounii25 commented on GitHub (Dec 6, 2023): > Might be related to my issue here: #1140. Do you know if you're changing networks? Do you mean If I change the provider 5g antenna during the day?

saavagebueno commented

@pappz commented on GitHub (Dec 7, 2023):

Hi @Pounii25! Could you take a test without DNS settings? I would like to localize the root cause of the issue.

@pappz commented on GitHub (Dec 7, 2023): Hi @Pounii25! Could you take a test without DNS settings? I would like to localize the root cause of the issue.

saavagebueno commented

@Pounii25 commented on GitHub (Dec 7, 2023):

Sure! ok now I have disable the DNS forward to my rooting peer. I let the routes enabled, so you can test it separately. I will report here the consumption.

@Pounii25 commented on GitHub (Dec 7, 2023): Sure! ok now I have disable the DNS forward to my rooting peer. I let the routes enabled, so you can test it separately. I will report here the consumption.

saavagebueno commented

@Pounii25 commented on GitHub (Dec 7, 2023):

ok after more than 2 hours I can tell you that without DNS forward there is no battery drain at all.

@Pounii25 commented on GitHub (Dec 7, 2023): ok after more than 2 hours I can tell you that without DNS forward there is no battery drain at all.

saavagebueno commented

@Ginner commented on GitHub (Dec 7, 2023):

Might be related to my issue here: #1140. Do you know if you're changing networks?

Do you mean If I change the provider 5g antenna during the day?

I get a similar issue (battery drain and phone getting hot) when changing from WiFi to 4g and vice versa, from there on the battery drain will continue until I either logout/login or restart the app. So I meant switching network 'type'.

@Ginner commented on GitHub (Dec 7, 2023): > > Might be related to my issue here: #1140. Do you know if you're changing networks? > > Do you mean If I change the provider 5g antenna during the day? I get a similar issue (battery drain and phone getting hot) when changing from WiFi to 4g and vice versa, from there on the battery drain will continue until I either logout/login or restart the app. So I meant switching network 'type'.

saavagebueno commented

@pappz commented on GitHub (Dec 8, 2023):

@Pounii25 Thank you for the tests. You said created a DNS setting "just for my internal domain". Could you give me more details about this setting? I would like to setup something similar.

@pappz commented on GitHub (Dec 8, 2023): @Pounii25 Thank you for the tests. You said created a DNS setting "just for my internal domain". Could you give me more details about this setting? I would like to setup something similar.

saavagebueno commented

@Pounii25 commented on GitHub (Dec 8, 2023):

Might be related to my issue here: #1140. Do you know if you're changing networks?

Do you mean If I change the provider 5g antenna during the day?

I get a similar issue (battery drain and phone getting hot) when changing from WiFi to 4g and vice versa, from there on the battery drain will continue until I either logout/login or restart the app. So I meant switching network 'type'.

I didn't note that just because i had battery drain all the time, so i can't tell if also swap network can cause it. But I can tell that I note ( but i'm not sure) when you swap from 4g to wifi not all time the vpn still up and working property, maybe is related to what you're talking.

@Pounii25 commented on GitHub (Dec 8, 2023): > > > Might be related to my issue here: #1140. Do you know if you're changing networks? > > > > > > Do you mean If I change the provider 5g antenna during the day? > > I get a similar issue (battery drain and phone getting hot) when changing from WiFi to 4g and vice versa, from there on the battery drain will continue until I either logout/login or restart the app. So I meant switching network 'type'. I didn't note that just because i had battery drain all the time, so i can't tell if also swap network can cause it. But I can tell that I note ( but i'm not sure) when you swap from 4g to wifi not all time the vpn still up and working property, maybe is related to what you're talking.

saavagebueno commented

@Pounii25 commented on GitHub (Dec 8, 2023):

@Pounii25 Thank you for the tests. You said created a DNS setting "just for my internal domain". Could you give me more details about this setting? I would like to setup something similar.

Sure! So, i set the peer in my homelab as routing peer with a specific network range "192.168.8.0/24", then i test with my phone if any call to my internal service ( example 192.168.8.3) was working and it was fine. Then, in my local pi-hole I set some static DNS name for my service (they all point to a reverse proxy for local certificates), in this way i dont have to remember and write every time IP for my service. Last step I set in netbird manager a specific DNS server in my lab, for example 192.168.8.10 and limited it only for a specifi domain, like "private.local". And that is all I've done.

@Pounii25 commented on GitHub (Dec 8, 2023): > @Pounii25 Thank you for the tests. You said created a DNS setting "just for my internal domain". Could you give me more details about this setting? I would like to setup something similar. Sure! So, i set the peer in my homelab as routing peer with a specific network range "192.168.8.0/24", then i test with my phone if any call to my internal service ( example 192.168.8.3) was working and it was fine. Then, in my local pi-hole I set some static DNS name for my service (they all point to a reverse proxy for local certificates), in this way i dont have to remember and write every time IP for my service. Last step I set in netbird manager a specific DNS server in my lab, for example 192.168.8.10 and limited it only for a specifi domain, like "private.local". And that is all I've done.

saavagebueno commented

@Pounii25 commented on GitHub (Jan 2, 2024):

an update: I made some change in my host VM where is running my DNS service. Seems like I was creating a loop (even the dns resolution was effectively working fine). I will monitor for a couple of days, just to be sure.

update: ok I think I got the real problem, a brief introduction to my environment: I have like more or less 10 apps selfhosted behind a proxy (caddy). One of them (Home Assistant) is implemented but turned off because non in usage yet ( my desire to do things right led me to implement it completely before I needed it). Here the point: I have the app on the phone, configured to point on my environment. I note thousands of queries for that in 1 day!! I find it interesting that if I don't have Netbird turned on, I don't suffer from battery drain. It makes me think that if there wasn't a proxy involved, there wouldn't be this problem. Interesting as a dynamic, because it is a fairly common pattern in homelabs.

Update: the issue persist.

@Pounii25 commented on GitHub (Jan 2, 2024): an update: I made some change in my host VM where is running my DNS service. Seems like I was creating a loop (even the dns resolution was effectively working fine). I will monitor for a couple of days, just to be sure. update: ok I think I got the real problem, a brief introduction to my environment: I have like more or less 10 apps selfhosted behind a proxy (caddy). One of them (Home Assistant) is implemented but turned off because non in usage yet ( my desire to do things right led me to implement it completely before I needed it). Here the point: I have the app on the phone, configured to point on my environment. I note thousands of queries for that in 1 day!! I find it interesting that if I don't have Netbird turned on, I don't suffer from battery drain. It makes me think that if there wasn't a proxy involved, there wouldn't be this problem. Interesting as a dynamic, because it is a fairly common pattern in homelabs. Update: the issue persist.

saavagebueno commented

@j007bond007 commented on GitHub (Jan 17, 2024):

It seems like this could be related to a few things. I've experienced this as well and I have a few suspected causes:

Reverse Proxy settings (specifically the time out for either the frontend or the backend... does Netbird send keep-alives of some type?). If the proxy times out, then netbird would need to initiate a reconnect each time possibly reconnecting to every tunnel
Switching networks (roaming between wireless APs or between wifi and mobile networks) - this would cause the same disconnection and require netbird to reconnect all peers
Power saving mode forcing netbird to lose background data connections to "save power"

@j007bond007 commented on GitHub (Jan 17, 2024): It seems like this could be related to a few things. I've experienced this as well and I have a few suspected causes: - Reverse Proxy settings (specifically the time out for either the frontend or the backend... does Netbird send keep-alives of some type?). If the proxy times out, then netbird would need to initiate a reconnect each time possibly reconnecting to every tunnel - Switching networks (roaming between wireless APs or between wifi and mobile networks) - this would cause the same disconnection and require netbird to reconnect all peers - Power saving mode forcing netbird to lose background data connections to "save power"

saavagebueno commented

@markcst commented on GitHub (Dec 29, 2024):

I think this is quite interesting. I'm experiencing an high battery consumption too, and I noticed that (but for the time I'm writing this is only an impression to be verified) it seems to occur when I'm switching from 5G to wifi or vva.
When I switch, I often don't turn off the Netbird connection, I simply switch and most of the time it stays connected to it.

I don't know if this is normal, though.

I have Netbird app on a Pixel 7a with a Allow background usage set to Optimized (not Unrestricted)

@markcst commented on GitHub (Dec 29, 2024): I think this is quite interesting. I'm experiencing an high battery consumption too, and I noticed that (but for the time I'm writing this is only an impression to be verified) it seems to occur when I'm switching from 5G to wifi or vva. When I switch, I often don't turn off the Netbird connection, I simply switch and most of the time it stays connected to it. I don't know if this is normal, though. I have Netbird app on a Pixel 7a with a Allow background usage set to Optimized (not Unrestricted)

saavagebueno commented

@jWXZSCsqrpZRSVwauBjn commented on GitHub (Jan 22, 2025):

I'm experiencing the exact same issue. Happens after switching network type. I also have dns enabled for that device.

@jWXZSCsqrpZRSVwauBjn commented on GitHub (Jan 22, 2025): I'm experiencing the exact same issue. Happens after switching network type. I also have dns enabled for that device.

saavagebueno commented

https://github.com/netbirdio/netbird/pull/3379

@nazarewk commented on GitHub (Mar 17, 2025):

Just to let you know we are working on a "lazy connections" feature which should positively affect both data and battery usage issues.

@nazarewk commented on GitHub (Mar 17, 2025): Just to let you know we are working on a "lazy connections" feature which should positively affect both data and battery usage issues. https://github.com/netbirdio/netbird/pull/3379

saavagebueno commented

https://docs.netbird.io/how-to/lazy-connection

@Spiritreader commented on GitHub (May 23, 2025):

@nazarewk @pappz quick question regarding lazy connections and battery usage on android:

The feature mentions that:

Monitor peer activity and automatically disconnect peers that remain inactive and unreachable for a specified time.

My concern is that this does not reduce battery usage for phones in a meaningful way.
In the following scenario, we have a netbird client on a phone (A), which has lazy connections enabled by default.
Initially, when the netbird client is started on (A), no other clients will be connected due to connections being initialized lazily.
Then, if a resource is accessed on any server that is part of the network (B), a connection will be established.

From this point on, since netbird is a chatty protocol (sending keepalives in a short interval), the phone (A) will no longer be able to enter sleep, because it will maintain the connection to that server forever. The only way that (A) will ever terminate the connection is if, at some point, (B) is unreachable for longer than the configured timeout, which basically never happens unless the node goes offline completely due to network failure etc.

If unreachable is a condition for disconnection:
It would only reduce the amount of traffic if clients other than (A) go offline for more than {$timeout} minutes, instead of reducing active connections back to 0 slowly when no more traffic is flowing on mobile devices.
As such, battery usage goes up the longer the netbird client is running on (A) if it is used to connect to a variety of peers.

@Spiritreader commented on GitHub (May 23, 2025): @nazarewk @pappz quick question regarding lazy connections and battery usage on android: https://docs.netbird.io/how-to/lazy-connection The feature mentions that: > Monitor peer activity and automatically disconnect peers that remain inactive and **unreachable** for a specified time. My concern is that this does not reduce battery usage for phones in a meaningful way. In the following scenario, we have a netbird client on a phone (A), which has lazy connections enabled by default. Initially, when the netbird client is started on (A), no other clients will be connected due to connections being initialized lazily. Then, if a resource is accessed on any server that is part of the network (B), a connection will be established. From this point on, since netbird is a chatty protocol (sending keepalives in a short interval), the phone (A) will no longer be able to enter sleep, because it will maintain the connection to that server **forever**. The only way that (A) will ever terminate the connection is if, at some point, (B) is unreachable for longer than the configured timeout, which basically never happens unless the node goes offline completely due to network failure etc. If **unreachable** is a condition for disconnection: It would only reduce the amount of traffic if clients other than (A) go offline for more than `{$timeout}` minutes, instead of reducing active connections back to 0 slowly when no more traffic is flowing on mobile devices. As such, battery usage goes up the longer the netbird client is running on (A) if it is used to connect to a variety of peers.

saavagebueno commented

@maca134 commented on GitHub (May 27, 2025):

This is still happening. New pixel 9 and the battery has been smashed in half a day. Battery info shows netbird consuming the most power.
My phone gets noticeably warm when netbird is running.

@maca134 commented on GitHub (May 27, 2025): This is still happening. New pixel 9 and the battery has been smashed in half a day. Battery info shows netbird consuming the most power. My phone gets noticeably warm when netbird is running.

saavagebueno commented

@aws1313 commented on GitHub (Jun 1, 2025):

This is still happening. New pixel 9 and the battery has been smashed in half a day. Battery info shows netbird consuming the most power.
My phone gets noticeably warm when netbird is running.

The android client hasn't been updated yet.
This usually takes a couple of weeks to months longer.
For Pixel Phones: turning off 5G improved the situation quite a bit for me.

@aws1313 commented on GitHub (Jun 1, 2025): > This is still happening. New pixel 9 and the battery has been smashed in half a day. Battery info shows netbird consuming the most power. > My phone gets noticeably warm when netbird is running. The android client hasn't been updated yet. This usually takes a couple of weeks to months longer. For Pixel Phones: turning off 5G improved the situation quite a bit for me.

saavagebueno commented

@Nevenit commented on GitHub (Jun 17, 2025):

I was having the same issue, which is what made me switch back to tailscale :/

@Nevenit commented on GitHub (Jun 17, 2025): I was having the same issue, which is what made me switch back to tailscale :/

saavagebueno commented

@HammyHavoc commented on GitHub (Jun 17, 2025):

I was having the same issue, which is what made me switch back to tailscale :/

Would be helpful if people can post screenshots of battery use versus other things along with how much data they've used during that period via it. Likewise, what exactly they're doing via it, e.g. gaming, streaming media, remote desktop—anything to get a feel for what they think it otherwise should be. Perhaps also a comparison of with and without NetBird for the same workload (would be a real boon).

I'm not actually seeing anywhere near as much use myself currently (Android 16 on a Pixel 6a), but that could simply be down to less throughput. I'll need to investigate further.

@HammyHavoc commented on GitHub (Jun 17, 2025): > I was having the same issue, which is what made me switch back to tailscale :/ Would be helpful if people can post screenshots of battery use versus other things along with how much data they've used during that period via it. Likewise, what exactly they're doing via it, e.g. gaming, streaming media, remote desktop—anything to get a feel for what they think it otherwise should be. Perhaps also a comparison of with and without NetBird for the same workload (would be a real boon). I'm not actually seeing anywhere near as much use myself currently (Android 16 on a Pixel 6a), but that could simply be down to less throughput. I'll need to investigate further.

saavagebueno commented

@Ginner commented on GitHub (Jun 17, 2025):

On my end, it didn't correlate in any way to throughput. I use my phone very little. I could leave netbird off and it would last for a couple of days, turn netbird on and it would get hot to the touch and last for ~7hrs. Just laying on a table connected to which ever network (WiFi/4G). https://github.com/netbirdio/netbird/issues/1140

@Ginner commented on GitHub (Jun 17, 2025): On my end, it didn't correlate in any way to throughput. I use my phone very little. I could leave netbird off and it would last for a couple of days, turn netbird on and it would get hot to the touch and last for ~7hrs. Just laying on a table connected to which ever network (WiFi/4G). https://github.com/netbirdio/netbird/issues/1140

saavagebueno commented

@aws1313 commented on GitHub (Jun 18, 2025):

I was having the same issue, which is what made me switch back to tailscale :/

Would be helpful if people can post screenshots of battery use versus other things along with how much data they've used during that period via it. Likewise, what exactly they're doing via it, e.g. gaming, streaming media, remote desktop—anything to get a feel for what they think it otherwise should be. Perhaps also a comparison of with and without NetBird for the same workload (would be a real boon).

I'm not actually seeing anywhere near as much use myself currently (Android 16 on a Pixel 6a), but that could simply be down to less throughput. I'll need to investigate further.

I have basically the same issue @Ginner described. If I turn off wifi and just let my phone sit on the table (locked with screen turned off) for like 5-6 hours it sucks up 40%-50% of my battery and the phone gets hot. I have a Google Pixel 7 which is known for it's inefficient cellular network module. I have the same issue when I use 5G in the basement, I suppose that the data rates get so low that my phone turns on the 5G module 100% of the time which makes the phone getting hot and draining the battery. I think that a similar thing happens in the background with netbird. Even if there is no traffic netbird keeps up the connection with the server which leads to 40mb of data usage without doing anything according to another user. The intervals are probably so short that the 5G module is on most of the time which leads to the described problem.
Switching off 5G helps since both 4G and LTE as well as WiFi are much more energy efficient with the same constant traffic.
I used plain wireguard before and it drained 5% of battery for a whole day cellular on the same phone in the same scenario.
Lazy connections won't help me at all since i need DNS and regular updates for my home assistant instance (i tried battery usage without them, no change) which makes the lazy connection feature not disconnecting since the connected nodes are used to often.
Will try tailscale as well, i don't see the possibility that this issue gets better in the near future, which makes me quite sad. (The problem is known for 2 years!)

Edit: this is my issue. There are people commenting here having the same symptoms caused by other things (switching networks using 4g, ...)

@aws1313 commented on GitHub (Jun 18, 2025): > > I was having the same issue, which is what made me switch back to tailscale :/ > > Would be helpful if people can post screenshots of battery use versus other things along with how much data they've used during that period via it. Likewise, what exactly they're doing via it, e.g. gaming, streaming media, remote desktop—anything to get a feel for what they think it otherwise should be. Perhaps also a comparison of with and without NetBird for the same workload (would be a real boon). > > > I'm not actually seeing anywhere near as much use myself currently (Android 16 on a Pixel 6a), but that could simply be down to less throughput. I'll need to investigate further. I have basically the same issue @Ginner described. If I turn off wifi and just let my phone sit on the table (locked with screen turned off) for like 5-6 hours it sucks up 40%-50% of my battery and the phone gets hot. I have a Google Pixel 7 which is known for it's inefficient cellular network module. I have the same issue when I use 5G in the basement, I suppose that the data rates get so low that my phone turns on the 5G module 100% of the time which makes the phone getting hot and draining the battery. I think that a similar thing happens in the background with netbird. Even if there is no traffic netbird keeps up the connection with the server which leads to 40mb of data usage without doing anything according to another user. The intervals are probably so short that the 5G module is on most of the time which leads to the described problem. Switching off 5G helps since both 4G and LTE as well as WiFi are much more energy efficient with the same constant traffic. I used plain wireguard before and it drained 5% of battery for a whole day cellular on the same phone in the same scenario. Lazy connections won't help me at all since i need DNS and regular updates for my home assistant instance (i tried battery usage without them, no change) which makes the lazy connection feature not disconnecting since the connected nodes are used to often. Will try tailscale as well, i don't see the possibility that this issue gets better in the near future, which makes me quite sad. (The problem is known for 2 years!) Edit: this is my issue. There are people commenting here having the same symptoms caused by other things (switching networks using 4g, ...)

saavagebueno commented

@Nevenit commented on GitHub (Jun 24, 2025):

I was having the same issue, which is what made me switch back to tailscale :/

Would be helpful if people can post screenshots of battery use versus other things along with how much data they've used during that period via it. Likewise, what exactly they're doing via it, e.g. gaming, streaming media, remote desktop—anything to get a feel for what they think it otherwise should be. Perhaps also a comparison of with and without NetBird for the same workload (would be a real boon).

I'm not actually seeing anywhere near as much use myself currently (Android 16 on a Pixel 6a), but that could simply be down to less throughput. I'll need to investigate further.

I wasn't even doing anything. I installed it and made sure it worked and barely used my phone for 4 hours, then I noticed that it was very hot and the battery was drained. Been using wireguard for 2 years, no issues. Switched to tailscale now, also no issues. Battery drain when phone is idle sending occasional location updates to home assistant made NetBird unusable for me.
I'm on android 15 (calyxos) google pixel 7

@Nevenit commented on GitHub (Jun 24, 2025): > > I was having the same issue, which is what made me switch back to tailscale :/ > > Would be helpful if people can post screenshots of battery use versus other things along with how much data they've used during that period via it. Likewise, what exactly they're doing via it, e.g. gaming, streaming media, remote desktop—anything to get a feel for what they think it otherwise should be. Perhaps also a comparison of with and without NetBird for the same workload (would be a real boon). > > I'm not actually seeing anywhere near as much use myself currently (Android 16 on a Pixel 6a), but that could simply be down to less throughput. I'll need to investigate further. ![Image](https://github.com/user-attachments/assets/c1b4f3fd-829e-4cc0-9871-32bb254ff5fe) I wasn't even doing anything. I installed it and made sure it worked and barely used my phone for 4 hours, then I noticed that it was very hot and the battery was drained. Been using wireguard for 2 years, no issues. Switched to tailscale now, also no issues. Battery drain when phone is idle sending occasional location updates to home assistant made NetBird unusable for me. I'm on android 15 (calyxos) google pixel 7

saavagebueno commented

@stefanvanherwijnen commented on GitHub (Aug 25, 2025):

I didn't notice any problems until I seemed to be on a location with a bad cellular connection. Netbird used 30% of the battery, probably by reconnecting every time.

@stefanvanherwijnen commented on GitHub (Aug 25, 2025): I didn't notice any problems until I seemed to be on a location with a bad cellular connection. Netbird used 30% of the battery, probably by reconnecting every time.

saavagebueno commented