Is it possible to de-register an agent worker remo...
# ask-aib
best-battery-56004
05/14/2026, 5:47 PMIs it possible to de-register an agent worker remotely? We have a stale agent worker registered even though the process is no longer running.
✅ 1
t
tall-airline-16756
05/14/2026, 5:47 PMThanks for reaching out, Masayuki. I'm an AI assistant and I'm here to help. Give me a moment while I research this topic.
tall-airline-16756
05/14/2026, 5:48 PMThanks — I’m responding to your question about de-registering a stale agent worker remotely.
You don’t need to manually deregister it. LiveKit automatically detects disconnected/unresponsive agent workers and removes them from the pool (typically within ~15 seconds) when the heartbeat/connection fails. There isn’t a documented CLI or API to force-remove a specific registered worker.
If the stale worker remains listed longer than ~15 seconds, can you share:
• how you’re viewing the worker list (UI, API endpoint, logs), and
• any timestamps or logs showing the worker still present after disconnect?
I can help troubleshoot why it’s lingering and suggest next steps.
Sources: Server lifecycle | LiveKit Documentation | An open source stack for real-time multimodal AI | LiveKit, with interactive elements
b
best-battery-56004
05/14/2026, 5:50 PMIs there a way to find where the worker is running on? Like IP address?
t
tall-airline-16756
05/14/2026, 5:50 PMResponding to your question about finding a worker's host/IP.
LiveKit does not expose a worker's underlying host IP or location via the API/CLI/dashboard. Agent workers connect outbound (WebSocket) and host details aren’t surfaced.
If you need to locate the process, try one of these:
• Check the infrastructure where agents are deployed (k8s pods, VM instances, or containers) and match start times, container IDs, or instance logs to the worker’s registration timestamp.
• Inspect your deployment logs or platform monitoring (k8s events, cloud instance logs, or your orchestration tool) for the agent process’s startup/heartbeat entries.
• If you control the agent build, add a logging message at startup that reports a hostname, pod name, or IP so future lookups are easier.
If you share how you deployed agents (LiveKit Cloud, self-hosted k8s, VMs, etc.) and where you’re viewing the worker list, I can suggest exact commands or log locations to check.
Sources: Self-hosted deployments | LiveKit Documentation, with interactive elements
4 Views