netbird
netbird copied to clipboard
POST /signalexchange.SignalExchange/Send API frequently
Describe the problem I use nginx proxy netbird signal, and got /signalexchange.SignalExchange/Send API call frequently when netbird android client disconnected
It seems that other netbird clients are still trying to exchange information with netbird client on android through signal server.
Expected behavior If the management service knows which clients are offline, can make other clients stop to call/signalexchange.SignalExchange/Send API for the offline android client?
NetBird status -d output: Daemon version: 0.24.3 CLI version: 0.24.3 Management: Connected to https://nb.xxx.com:8443 Signal: Connected to https://nb.xxx.com:8443 FQDN: m1.nb.iot NetBird IP: 100.64.0.3/16 Interface type: Kernel Peers count: 3/4 Connected
Hello @xiaolei0125, this is the current behavior of the client. We have plans to optimize this in the next quarter and prevent constant attempts to connect, even if the peer is offline.
Any update?
+1 we still got this prob with version v0.27.4
@bcmmbaga I have created netbirds on many devices, but most of them are offline most of the time. I have received a large number of requests for these offline devices, and I think this is unreasonable.
@bcmmbaga I have created netbirds on many devices, but most of them are offline most of the time. I have received a large number of requests for these offline devices, and I think this is unreasonable.
That's true, we have 150 servers, when only a few of them got offline, the caddy and signal service used almost 100% CPU usage of a 4C8G server.
Any update on this? We have 2500 servers planned to install. This sounds scary?
still got this prob with version v0.28.2
@bcmmbaga @mlsmaycon @lixmal
Do you have a plan when to solve this problem?
I think when the number of devices is large, this problem becomes very obvious.
The signal reconnection shouldn't be an issue for connectivity as it happens automatically and it only does that, it shouldn't interrupt any existing peer connection.
We are looking at optimizing network map updates as now it is just brute force as it sends to all peers all the time there is an update even if there is no update for the peer.