Hey there! You can easily check out connection minutes and bandwidth details for a session in LiveKit using a few handy methods: • LiveKit Cloud Analytics Dashboard: This dashboard gives you a clear view of session-level metrics, like total connection minutes and bandwidth usage. Plus, you can look at data over specific time ranges to get the insights you need. • Analytics API: If you prefer, you can use the API to grab detailed session info. For example, just make a GET request to this endpoint:

<BASE_API_URL>/api/project/{PROJECT_ID}/sessions/{SESSION_ID}

. • Programmatic Access via SDKs: Another option is the

get_rtc_stats()

method in the LiveKit RTC SDKs. It helps you collect WebRTC stats, including publisher and subscriber info, which is great for figuring out bandwidth usage and connection minutes. If you want me to walk you through any of these or need more details, just let me know—I’m here to help! Sources: Say hello to improved Cloud analytics | livekit-rtc/livekit/rtc/room.py | livekit-ffi/src/server/requests.rs | packages/livekit-rtc/src/proto/room_pb.ts | Say hello to improved Cloud analytics

Tell me more about the below. • Programmatic Access via SDKs: Another option is the

get_rtc_stats()

method in the LiveKit RTC SDKs. It helps you collect WebRTC stats, including publisher and subscriber info, which is great for figuring out bandwidth usage and connection minutes. Can I call it from agent? or how to get this

Absolutely! You can call the

get_rtc_stats()

method from an agent in LiveKit. This handy method is part of the Room object and gives you RTC statistics, including both publisher and subscriber stats. These stats are super useful for figuring out bandwidth usage and connection minutes. Here’s a quick example to show you how it works:

async def entrypoint(ctx: JobContext):
    session = AgentSession(llm=openai.realtime.RealtimeModel())
    await session.start(Agent(instructions="You are a helpful assistant"), room=ctx.room)

    # Log the session stats every 5 minutes
    while True:
        rtc_stats = await ctx.room.get_session_stats()

        all_stats = chain(
            (("PUBLISHER", stats) for stats in rtc_stats.publisher_stats),
            (("SUBSCRIBER", stats) for stats in rtc_stats.subscriber_stats),
        )

        for source, stats in all_stats:
            stats_kind = stats.WhichOneof("stats")
            <http://logger.info|logger.info>(f"RtcStats - {stats_kind} - {source}", extra={"stats": MessageToDict(stats)})

        await asyncio.sleep(5 * 60)

Just a heads-up: make sure your room is connected before calling this method, because it will raise a RuntimeError if it’s not. If you have any questions or want me to walk you through it, just let me know—I’m here to help!