LiveKit Community

Separately, is there a way to set the volume of the track? I don't see that as an option in TrackPublicationOptions

<@U01NZKTNE1H> any clue? seems to happen on reconnections?

sorry, haven't gotten around to digging deep on this one, wondering if it's some sort of race condition happening during reconnect

tracking here: <https://github.com/livekit/client-sdk-android/issues/110>

I noticed another crash <@U01NZKTNE1H>
```Fatal Exception: java.lang.NullPointerException: Attempt to invoke virtual method 'java.lang.Class java.lang.Object.getClass()' on a null object reference
       at livekit.LivekitRtc$SyncState.setAnswer(LivekitRtc.java:17847)
       at livekit.LivekitRtc$SyncState.access$35600(LivekitRtc.java:17818)
       at livekit.LivekitRtc$SyncState$Builder.setAnswer(LivekitRtc.java:18221)
       at io.livekit.android.room.RTCEngine.sendSyncState(RTCEngine.kt:754)
       at io.livekit.android.room.Room.sendSyncState(Room.kt:446)
       at io.livekit.android.room.Room.onSignalConnected(Room.kt:667)
       at io.livekit.android.room.RTCEngine$reconnect$job$1.invokeSuspend(RTCEngine.kt:383)
       at kotlin.coroutines.jvm.internal.BaseContinuationImpl.resumeWith(ContinuationImpl.kt:33)
       at kotlinx.coroutines.DispatchedTask.run(DispatchedTask.kt:106)
       at kotlinx.coroutines.internal.LimitedDispatcher.run(LimitedDispatcher.kt:42)
       at kotlinx.coroutines.scheduling.TaskImpl.run(Tasks.kt:95)
       at kotlinx.coroutines.scheduling.CoroutineScheduler.runSafely(CoroutineScheduler.kt:570)
       at kotlinx.coroutines.scheduling.CoroutineScheduler$Worker.executeTask(CoroutineScheduler.kt:749)
       at kotlinx.coroutines.scheduling.CoroutineScheduler$Worker.runWorker(CoroutineScheduler.kt:677)
       at kotlinx.coroutines.scheduling.CoroutineScheduler$Worker.run(CoroutineScheduler.kt:664)```

Hey, <@U03MFML4QH1>, this seems similar to this issue?

<https://livekit-users.slack.com/archives/C01KVTJH6BX/p1657047942356519|https://livekit-users.slack.com/archives/C01KVTJH6BX/p1657047942356519>

<@U02MAJLGN65> 

Hi <@U03MFML4QH1>, I’ll try to get this working.
(related to AdaptiveStream feature)

Hi, improvements (multiple VideoViews per track) have been merged into `main` branch of the SDK.
I’ll try to get more fixes in and tag a new version.

this is likely a bug, could you file an issue in the react-native repo?

Hi <@U03C8KQDJMA>,
that’s correct. the capture defaults are the default settings with which LiveKit will try to acquire local media tracks from the browser. Publish defaults are the settings for how these tracks should be sent to the LiveKit server

swift package manager can take a branch or specific commit hash:

<https://developer.apple.com/documentation/packagedescription/package/dependency/requirement-swift.enum/branch(_:)>

for example:

```.package(name: "LiveKit", url: "<https://github.com/livekit/client-sdk-swift.git>", .branch("main")),```


the latest swift example with swift sdk v1.0.1 works for me, tested against server version v1.1.1

can you confirm which versions you're using?

Sorry guys we are still checking the problem but is something related with our deployment. Was weird because iOS was down but the others clients where all ok. Now also using the official web demo is giving us an "error could not connect after timeout"

that version did have a change that impacted connectivity for some environments.. we are releasing 1.1.1 today

Ok we do a fresh server install (latest 1.1.1) and everything is working fine :+1:

Hello <@U03NBU0MGA1>, I would say for an example, the one which the LiveKit built would be a great starting point:

Demo: <https://example.livekit.io/>
Source: <https://github.com/livekit/livekit-react/tree/master/example>

It's still hard for me <@U032TUAL4KT>, lot of magic (like react component dependencies)

edit:

I made it slowly

great news. If you are not familiar with React, maybe starting with the JS version first might be easier.

The example is here: <https://github.com/livekit/client-sdk-js/tree/main/example> .. It has a lot of options available, but you can slim those down as needed.

Hi, thanks for the help, I already made my first live kit app

and currently I'm working on Twitter Space / Clubhouse clone

Well that's awesome! :slightly_smiling_face: Happy to hear that.

if you want to manually switch, you can pass in an `AudioSwitchHandler` to LiveKitOverrides and use the `selectDevice` method to select which device to output

<https://github.com/livekit/client-sdk-android/blob/main/sample-app-common/src/main/java/io/livekit/android/sample/CallViewModel.kt#L84>

Can you provide me the latest code for this solution

ah it might not be released yet, I'll cut a release in the next day or so

Hey <@U01NZKTNE1H> can you provide me the old code to change the sound output. In this new version its hard to understand the while code quickly for me.

There is so much change in 0.9.1 and 1.1.0

the current version is the only version that has had audio output selection

Its not possible to change sound output in the 0.9.1 version?

You would have to do that yourself through android

I just want to set the default output to the device speaker. Nothing else.

An AudioSwitchHandler can be provided to LiveKit overrides, you can set the default preferred output there

Can you please tell me the code to set speaker to the default. I used your default code still the audio output is from earpiece.

Try like this:
```private val preferredDeviceList = listOf(
    AudioDevice.BluetoothHeadset::class.java,
    AudioDevice.WiredHeadset::class.java,
    AudioDevice.Speakerphone::class.java,
    AudioDevice.Earpiece::class.java,
)

            audioHandler.preferredDeviceList = preferredDeviceList```

It is possible currently to protect LiveKit video tag with DRM? (like what netflix does)

hi <@U03NYRNS9JN>, are you trying to connect to the same room twice with the same participant identity (i.e. access token)?

I don't think so...  I just plug the token I get from `docker run --rm -v$PWD:/output livekit/generate --local`
into the web form at <https://example.livekit.io/>

Do you think you can share the full server logs with me? that does seem strange.

Server version: `1.1.0`, Client version: `1.0.1`

android client correct? this seems like the WebRTC side couldn't establish an ice connection.

would you happen to have server/client logs? particularly looking to see if the server sent the information regarding ice candidates and offer/answer

I debugged on android, but iOS wasn't connecting either, so it was definitely a server-side issue (and also because rebooting the server fixed).
I'll ask out backend team to check if they have any logs

do you know what substring we should be looking for in the logs?

these are my logs from joining a room, looking for things like "sending server offer" "ice candidates"

participant state: `ACTIVE` would mean that they've established ice connection

image.png

also no messages about ice candidates in the log

but there are no messages about "ice candidates" now either. even though the connection works

nothing. neither when we had the problem, nor now

basically that's all I have now in the logs for a successful connection:

Does the `stuck` mean all clients joined from that point are stuck at that state?

yea clients never got more messages through the websocket after the first one and never transitioned to a following state (and they didn't timeout either). actually when we rebooted the client unlocked, so closing the websocket did trigger a callback.

can you post a entire log that a client joined after server go to `stuck` state

```Jul 8, 2022 @ 12:42:21.811{"level":"debug","ts":1657284141.8115067,"logger":"layer4","msg":"connection stats","remote":"78.61.255.178:38903","read":547,"written":4845,"duration":0.042620814}
Jul 8, 2022 @ 12:42:21.777{"level":"debug","ts":1657284141.7770457,"logger":"tls.handshake","msg":"choosing certificate","identifier":"<http://1.lk.de-turn.jaumo.com|1.lk.de-turn.jaumo.com>","num_choices":1}
Jul 8, 2022 @ 12:42:21.777{"level":"debug","ts":1657284141.777081,"logger":"tls.handshake","msg":"default certificate selection results","identifier":"<http://1.lk.de-turn.jaumo.com|1.lk.de-turn.jaumo.com>","subjects":["<http://1.lk.de-turn.jaumo.com|1.lk.de-turn.jaumo.com>"],"managed":true,"issuer_key":"acme-v02.api.letsencrypt.org-directory","hash":"a913f7e414ea697652463e801f63e36b181e5533c9118805bac50883d9b4792e"}
Jul 8, 2022 @ 12:42:21.777{"level":"debug","ts":1657284141.7770925,"logger":"tls.handshake","msg":"matched certificate in cache","subjects":["<http://1.lk.de-turn.jaumo.com|1.lk.de-turn.jaumo.com>"],"managed":true,"expiration":1664700777,"hash":"a913f7e414ea697652463e801f63e36b181e5533c9118805bac50883d9b4792e"}
Jul 8, 2022 @ 12:42:21.776{"level":"debug","ts":1657284141.7767467,"logger":"layer4.matchers.tls","msg":"matched","remote":"78.61.255.178:38903","server_name":"<http://1.lk.de-turn.jaumo.com|1.lk.de-turn.jaumo.com>"}
Jul 8, 2022 @ 12:42:21.4122022-07-08T12:42:21.412Z DEBUG livekit rtc/participant.go:725 SetMigrateState {"room": "b993655d-9368-4022-8272-28d921e5f2fa", "roomID": "RM_m9kskqTk2k3k", "participant": "276037971", "pID": "PA_yhEaVagyYC7W", "state": "MIGRATE_STATE_COMPLETE"}
Jul 8, 2022 @ 12:42:21.4112022-07-08T12:42:21.411Z DEBUG livekit rtc/participant.go:1057 updating participant state {"room": "b993655d-9368-4022-8272-28d921e5f2fa", "roomID": "RM_m9kskqTk2k3k", "participant": "276037971", "pID": "PA_yhEaVagyYC7W", "state": "JOINED"}
Jul 8, 2022 @ 12:42:21.4112022-07-08T12:42:21.411Z DEBUG livekit rtc/room.go:228 participant state changed {"room": "b993655d-9368-4022-8272-28d921e5f2fa", "roomID": "RM_m9kskqTk2k3k", "state": "JOINED", "participant": "276037971", "pID": "PA_yhEaVagyYC7W", "oldState": "JOINING"}
Jul 8, 2022 @ 12:42:21.4102022-07-08T12:42:21.409Z DEBUG livekit service/roommanager.go:227 starting RTC session {"room": "b993655d-9368-4022-8272-28d921e5f2fa", "nodeID": "ND_81LNDS9q", "participant": "276037971", "sdk": "ANDROID", "sdkVersion": "1.0.0", "protocol": 7}
Jul 8, 2022 @ 12:42:21.4102022-07-08T12:42:21.410Z INFO livekit rtc/room.go:276 new participant joined {"room": "b993655d-9368-4022-8272-28d921e5f2fa", "roomID": "RM_m9kskqTk2k3k", "pID": "PA_yhEaVagyYC7W", "participant": "276037971", "protocol": 7, "options": {"AutoSubscribe":true}}
Jul 8, 2022 @ 12:42:21.4072022-07-08T12:42:21.406Z INFO livekit service/rtcservice.go:211 new client WS connected {"room": "b993655d-9368-4022-8272-28d921e5f2fa", "participant": "276037971", "connID": "CO_r5vaSSgnPn6z", "roomID": "RM_m9kskqTk2k3k"}
Jul 8, 2022 @ 12:42:21.4032022-07-08T12:42:21.403Z DEBUG livekit service/roomallocator.go:107 selected node for room {"room": "b993655d-9368-4022-8272-28d921e5f2fa", "roomID": "RM_m9kskqTk2k3k", "nodeID": "ND_81LNDS9q"}
Jul 8, 2022 @ 12:42:21.288{"level":"debug","ts":1657284141.287877,"logger":"layer4.handlers.proxy","msg":"dial upstream","remote":"78.61.255.178:44348","upstream":"localhost:7880"}
Jul 8, 2022 @ 12:42:21.287{"level":"debug","ts":1657284141.2871687,"logger":"layer4.handlers.tls","msg":"terminated TLS","remote":"78.61.255.178:44348","server_name":"<http://1.lk.de.jaumo.com|1.lk.de.jaumo.com>"}
Jul 8, 2022 @ 12:42:21.158{"level":"debug","ts":1657284141.1575992,"logger":"tls.handshake","msg":"choosing certificate","identifier":"<http://1.lk.de.jaumo.com|1.lk.de.jaumo.com>","num_choices":1}
Jul 8, 2022 @ 12:42:21.158{"level":"debug","ts":1657284141.1576846,"logger":"tls.handshake","msg":"default certificate selection results","identifier":"<http://1.lk.de.jaumo.com|1.lk.de.jaumo.com>","subjects":["<http://1.lk.de.jaumo.com|1.lk.de.jaumo.com>"],"managed":true,"issuer_key":"acme-v02.api.letsencrypt.org-directory","hash":"5d28df81b4a0fdf91cf46c98c6a31f1fa926316a2d15d0e615656fcdf05fd289"}
Jul 8, 2022 @ 12:42:21.158{"level":"debug","ts":1657284141.157715,"logger":"tls.handshake","msg":"matched certificate in cache","subjects":["<http://1.lk.de.jaumo.com|1.lk.de.jaumo.com>"],"managed":true,"expiration":1664700777,"hash":"5d28df81b4a0fdf91cf46c98c6a31f1fa926316a2d15d0e615656fcdf05fd289"}
Jul 8, 2022 @ 12:42:21.157{"level":"debug","ts":1657284141.1569836,"logger":"layer4.matchers.tls","msg":"matched","remote":"78.61.255.178:44348","server_name":"<http://1.lk.de.jaumo.com|1.lk.de.jaumo.com>"}```

no offer created, looks strange… no more lines for the log ?

nope, at least for that user. but nothing about "offer created" at this moment either, while it works fine

`sending server offer to participant` it should have this line if it can work

I will check our logging, maybe something is filtered out there

also, can you enable the pion log during your test. in your config.yaml,
```logging:
    pion_level: info```

ok, we don't have this line now, because we set the level to `warning`, as the logs were consuming too much traffic
BUT we did not have it while the clients were hanging for real
after restarting the server, this line started appearing again

so the logs that I sent for the "hanging" period are complete - that's all we had at that moment

It would be great if it is possible to set the server log level to `debug` at least while doing this test (or `info` if `debug` is too noisy, but would be great to get debug logs at least to understand this issue).

Also, if you have try latest on `master` , that would be great too.

Relevant config section - <https://github.com/livekit/livekit/blob/4280a5993b6d5b682d7ce598afe1cc1294d4643f/config-sample.yaml#L95>

Setting `level: debug` in addition to the Pion log level <@U02GP9AENTH> requested above and providing the resulting logs would be great.

<https://livekit-users.slack.com/archives/C01KVTJH6BX/p1657536468033429?thread_ts=1657524136.395759&amp;cid=C01KVTJH6BX> - these are DEBUG logs

Thank you <@U03NU1ZARQS>. A few more questions
• How many clients are able to connect before the server starts hanging?
• If you have full logs of clients connecting and then the point at which server starts hanging, would be great to see the logs around the transition point.
• When the server is stuck, are all the clients connected at that point continue to function as expected? i. e. able to send/receive media.
• When the server is stuck, can the already connected clients hang up successfully?

&gt;  How many clients are able to connect before the server starts hanging?
difficult to say. we had the server running for 2 weeks. we are not even sure that the problem started because of high client numbers, we actually noticed it while there were very few clients active (but we had a big event few hours ago)
&gt;  If you have full logs of clients connecting and then the point at which server starts hanging, would be great to see the logs around the transition point.
unfortunately we  don't, as we don't know the exact time when it started happening
&gt;  When the server is stuck, are all the clients connected at that point continue to function as expected?
also unknown, sadly
&gt;  When the server is stuck, can the already connected clients hang up successfully?
no, there are no "closing participant", "client leaving room", "source closed connection" in the logs
basically everything ends on `MIGRATE_STATE_COMPLETE`

<@U03NU1ZARQS> Can you enable pion log then test again? it will  provide more details for the stuck

yeah, we enabled pion, will provide the logs when it happens again

There is a CORS error in the console ... maybe that's causing this.

Certainly didn't used to have this problem.

Hi <@U02HS7BES0H>,
could you share the CORS error that you get?
The participant should not get disconnected just by setting screen share to false…
Does the same thing happen also for Microphone/Camera or is it just ScreenShare?

<@U02EU2ZPTQU> This is what I see in the console. I've masked my domain name but if you need more info about the errors, please let me know.

Let me try disabling mic/camera ... by the way, none of this used to happen, but I'm trying to migrate to a new version of livekit and I'm using some APIs that didn't used to exist (such as creating a room and configuring it before connecting).

Ah for me turning off the mic or camera I'm using a different API, which is calling `.mute` on the track, which appears to work fine and has no weird errors like this.

Ok, I just tested it by calling the functions directly on the `local-participant` object, but I *do not* have this problem at all when I call `setCameraEnabled` or `setMicrophoneEnabled` (true or false).

thanks for the additional info, just to go step by step in figuring this out, it’d be great if you could go to:
1. <https://livekit.io/connection-test> and post the results
2. test in your browser on <https://example.livekit.io>, try screen sharing and see if it behaves in the same way

Interesting. For #1, everything passes *EXCEPT* connecting via TURN. (I wonder how to debug this?)

that seems ok though, TURN isn’t needed for it to work

So interestingly, when I start/stop screen sharing from `<http://example.livekit.io|example.livekit.io>`, I experience no problems.

I wonder what API is being used there ...

I mean an obvious consideration is that I'm connected to localhost via HTTP (as opposed to HTTPS) in my development environment and perhaps that could be causing these CORS problems, even though historically I've had no problem.

the example source code is public and also using `setScreenShareEnabled`.
Could it be that what you’re seeing is a difference in a newer brave version?

No, I'm almost certain that's not the case.

Anyway, thank you for your attention and help. I wonder if I'd have any better luck if I switched to the react SDK.

If you continue to run into troubles, I can take a look at it if you create a minimal reproduction on something like stackblitz.
You can also PM me the stackblitz project if you don’t want to share it in the public channel.

<@U02EU2ZPTQU> That's very kind of you. I'll have to see if that's something I can do.

<@U02EU2ZPTQU> Any tips on debugging why a TURN connection won't work?

The connection tester just says it timed out.

verify that your ports are set up correctly and that you defined TURN domains in your charts/deploy script accordingly (if using TCP)
<https://docs.livekit.io/deploy/>
<https://docs.livekit.io/deploy/ports-firewall/>

<@U02EU2ZPTQU> Is there any way to verify that livekit is actually running its turn server? There's no indication in the logs I can find, and I can't connect to that port directly with `nc -v` ... so I'm wondering if the reason I can't use TURN is ... it's not running.

as long as you’ve enabled it in your config, it should be running. try with udp first, as that’s easier to setup (doesn’t require certificates): <https://docs.livekit.io/deploy/#turn/udp>

The whole problem here is that I'm running into a bug in Safari related to UDP and VPN.

I already have certs in place, and I see this log message when I start up:

```Jul 18 12:16:03 localhost bash[611811]: 2022-07-18T12:16:03.480Z#011INFO#011livekit#011service/turn.go:112#011Starting TURN server#011{"turn.relay_range_start": 30000, "turn.relay_range_end": 30002, "turn.portTLS": 5349, "turn.externalTLS": false, "turn.portUDP": 3478}```
But interestingly, it doesn't appear to listening on 5349, at least not according to `netstat -l` .

In any case, thank you - maybe getting turn working is a lost cause, and it shouldn't be the root of the issue anyway. I just wanted see if it having a turn server happened to work around this browser bug.

Sorry, this must be my fault - it's because I'm an idiot related to docker.

Were you able to confirm that TURN/UDP is working? (if not with Safari, then with another browser)

<@U02EU2ZPTQU> So I've got TURN/UDP working in chrome but not in Safari in my configuration. I'm hoping to get TURN working over TCP, and I have the domain names and certs (I think) that I need, but I never see any output on the server (or the client, for that matter) regarding TURN. Is there any way to get more informative logging for what is happening? I turned the pion logging all the way up but I don't see mention of TURN or  the TURN ports even in the logs.

The connection tester just says it couldn't connect after a timeout.

can you share your livekit config part that configures TURN?

Yeah but I'm not sure it will be very illuminating due to apache being in front of everything. One moment.

I've obscured my specific domain, but it's basically this:

```# turn server
turn:
   enabled: true
   tls_port: 5349
   external_tls: true
   domain: <http://my-turn.mydomain.com|my-turn.mydomain.com>```

<@U02EU2ZPTQU> I have tried not terminating TLS with apache but the certs are causing livekit to freak out, probably because they require a password, so I will generate a new cert later with no password and try that to take apache out of the equation.

Does my turn server need to be visible to the (google) stun server?

(I ask because my livekit instance is on an internal corporate network)

I've decrypted the TLS traffic enough to verify that my turn server info is being returned to the caller (with port 443 being specified, which seems correct due to apache) ...

hm, your setup sounds tricky. do you even need TCP TURN if the instance is only running in a controlled internal network?

Well, I'm still trying to address a specific incompatibility with Safari in certain corporate VPN configurations. I was hoping if I could actually get TURN over TCP working, that it might be a workaround, since otherwise this configuration seems to not work with Safari.

It works fine without turn with other browsers, and also with Safari if I change the VPN configuration to force a full tunnel, but I can't really expect users to change their VPN config in order to use webRTC,

so I was just hoping to find some other workaround by seeing if TURN/TCP helped.

I'm not sure my network setup is that weird, except that there are hosts that aren't visible on the public internet.

What's odd to me is that no debugging information is ever emitted like "trying to connect to host xyz" in a network application like this.

<@U02EU2ZPTQU> I also see no evidence in the access logs from apache that anything actually tries to connect to my turn server.

I will try turning up apache logs and see if anything interesting happens.

yeah, this is the <https://docs.livekit.io/client-sdk-js/modules.html#AdaptiveStreamSettings|default behavior> of adaptive stream. you can disable it with
```adaptiveStream: {
  pauseVideoInBackground: false
}```


I have a different ssl certificate in place for turn server.

Hi, is your goal to switch front / back cameras ?
There is no need to re-publish, we just need to tell the `CameraCapturer` to restart with with new options.

`CameraCapturer.switchCameraPosition()` :
<https://github.com/livekit/client-sdk-swift/blob/main/Sources/LiveKit/Track/Capturers/CameraCapturer.swift#L54-L64>

Sample code calling `switchCameraPosition`:
<https://github.com/livekit/client-sdk-swift/blob/main/Sources/LiveKit/SwiftUI/ObservableRoom.swift#L49-L60>

There is also a `CameraCapturer.canSwitchPosition()`
if you would like to check if it’s switchable.
<https://github.com/livekit/client-sdk-swift/blob/main/Sources/LiveKit/Track/Capturers/CameraCapturer.swift#L31-L36>

in your case maybe something like:
```if let cameraCapturer = videoTrack.capturer as? CameraCapturer {
  cameraCapturer.switchCameraPosition()
}```

There is also `setCameraPosition` if you want to specify `.front` or `.back`

you should get a `RoomEvent.Disconnected` event when that happens. The latest server version passes an argument for that back to the client with a `DisconnectReason`.
`DisconnectReason.ParticipantRemoved` is the one you are after.

you can enable debug logs with `setLogLevel('debug')` which prints all room events to the console

you don't need to set videoCodec here. Egress will need to transcode to something suitable for storage.

honestly you'd only need to set resolution to 1080p, so that captured video remains sharp

<@U01KZ61GV5Z>
Thanks.
Maybe there is a little misunderstanding because the last settings like `videoCaptureDefaults` are client-side `RoomOptions`.
I thought it's a good idea to synchronise client-side and Egress as much as possible to deliver best quality and avoid heavy transcoding.

gotcha. I assume you are using track composite egress?

it's a great idea to automatically egress the same bitrate/resolution you are publishing at. unfortunately we'd still have to transcode, because there are differences between media encoded for streaming and media encoded for storage

<@U01KZ61GV5Z>
I use track composite, yes!

Okay: I can't avoid CPU load for transcoding.
*But at least I ensure or limit the quality delivered for recording by setting `RoomOptions` on client-side this way, right?*

If that's the case, I'd like to know what are the matching values?

*Example 1:*

On client-side there `TrackPublishDefaults` like `audioBitrate`.
It seems that in contrast to the server-side Egress option with same property name, it's not in `kbit/s` but `bit/s` on client-side.

*Example 2:*

On client-side there are `AudioCaptureOptions` like `sampleRate` and `sampleSize` - on Egress I use
```audioFrequency: 48_000,```
Is `sampleRate` the `audioFrequency`?

yes.. that makes sense.. client takes it in bits currently.. that. inconsistency is not great.

and yes, sampleRate is audioFrequency.

Oh my bad.. I should have picked Startup Shell Script option at the beginning. Sorry!

oops, that line is out of date.. they have been moved to `localParticipant.lastCameraError` , etc

<https://github.com/livekit/client-sdk-js/blob/main/src/room/participant/LocalParticipant.ts#L97>

Ok, thanks! I just opened a PR which updates the docs accordingly.

thanks for the report Jorn! it's a regression on our side and we'll investigate

Hey. I've been trying to get metrics, but it was to no avail. The steps I've gone through:
1. generate yaml for aws
2. set `prometheus_port: 6789` in the content of /opt/livekit/livekit.yaml in amazon.yaml file
3. launched an instance with the that file as userdata
4. opened the port to public via security group
So I tried to just request the resource "domain:6789/metrics" and I get nothing. No idea what I am supposed to do to make it work. Any help would be appreciated

are `room.NumParticipants` from `roomClient.ListRooms` not realtime updated ?

```ERROR livekit routing/redisrouter.go:261 could not handle new participant	{"room": "7fa3aa36-120e-427c-bf5c-1b6a0ab0317d", "participant": "511178735", "error": "could not restart participant", "errorVerbose": "could not restart participant\<http://ngithub.com/livekit/livekit-server/pkg/service.(*RoomManager).StartSession\n\t/workspace/pkg/service/roommanager.go:225\ngithub.com/livekit/livekit-server/pkg/routing.(*RedisRouter).startParticipantRTC.func1\n\t/workspace/pkg/routing/redisrouter.go:253\nruntime.goexit\n\t/usr/local/go/src/runtime/asm_amd64.s:1581%22}|ngithub.com/livekit/livekit-server/pkg/service.(*RoomManager).StartSession\n\t/workspace/pkg/service/roommanager.go:225\ngithub.com/livekit/livekit-server/pkg/routing.(*RedisRouter).startParticipantRTC.func1\n\t/workspace/pkg/routing/redisrouter.go:253\nruntime.goexit\n\t/usr/local/go/src/runtime/asm_amd64.s:1581"}>```

hi <@U03PRK0N1QU>
we have a feature built in for that: <https://docs.livekit.io/guides/room/receive/#adaptive-stream>
This will automatically attach listeners to the video element and set the subscribed simulcast quality accordingly to its size

Oh that's great, thank you <@U02EU2ZPTQU>

Hello, trying to install LK cli w/ `curl -sSL <https://get.livekit.io/cli> | bash` but getting an error `bash: line 48: /usr/bin/uname: No such file or directory` This is on a jetson Xavier w/ latest L4T

hm nm, might be something on my machine, /usr/bin/uname isn’t there…

you'd call <https://docs.livekit.io/guides/server-api/#listrooms|ListRoom>, filtering by room name. It will include a field <https://github.com/livekit/protocol/blob/main/livekit_models.proto#L19|num_participants>

Sorry, but I find It's not real-time updated, Like I need to wait until 2/more minutes to get the room participant updated

```roomClient.
			ListRooms(context.Background(), &amp;livekit.ListRoomsRequest{})```