Is LiveKit down?

A subset of room sessions that ended around 4 September 09:57 UTC were not recorded with an end time. These sessions continue to appear as active in the Cloud dashboard and in session records, and no `room_ended` end time is reflected for them. This was a reporting issue only. Rooms ended normally for participants, live traffic was not affected, and there is no impact on usage or billing. The underlying cause has been identified and fixed. Affected sessions from this window may continue to display as active; they can be safely disregarded.

We have identified an issue with LiveKit Cloud hosted agents in eu-central. Existing agents in eu-central are facing scaling issues resulting in increased join latencies. Existing agents are still functional and can serve sessions normally, but they may be slower to join new sessions. Additionally, new agents cannot be created in this region. We are actively working on a mitigation to alleviate this issue. We will provide another update as soon as possible.

We have applied a mitigation and believe that symptoms should be improving. We will update again with confirmation as soon as possible.

The mitigation was successful and users shouldn't notice any issues with join latencies or creating new agents. We will follow up as soon as possible with a postmortem.

**Root Cause** A configuration change introduced a formatting error in the startup settings for the server pools that run hosted agents in our eu-central region. Newly provisioned servers failed to start, so the region could not add capacity. Agents that were already running were unaffected, but new agent deployments, version rollouts, and automatic scale-ups were delayed or stuck pending until the change was reverted. **Timeline \(UTC\)** `2026-09-04 23:56` - Configuration change deployed. The error only affected newly provisioned servers, so there was no immediate impact. `2026-09-05 18:07` - New servers began failing to start. Impact begins, intermittent at first. `2026-09-05 22:23` - Our monitoring alerted us to agent deployments stuck pending and we began investigating. `2026-09-06 00:50` - We identified the malformed startup configuration and reverted the change. `2026-09-06 01:17` - New capacity provisioned successfully, all pending deployments recovered, and we validated the fix. **Scope of Impact** Limited to hosted agents in our eu-central region. During the impact window, new agent deployments and rollouts were delayed or stuck pending, and automatic scale-up was blocked. Agents already running continued to serve sessions normally, and we estimate that, at the peak, roughly 1% of agent instances in the region were affected. No other regions or products were affected. **Mitigations and Follow-ups** * The faulty configuration change was reverted and provisioning has been stable since. * We are adding alerting for new servers that fail to start via this failure mode, which today fails silently. This lets us detect this class of failure directly and validate future changes to server provisioning quickly. We appreciate your understanding and are committed to continuously improving our platform's reliability. If you have any questions, please reach out to our support team.

We are investigating elevated issues in calls connected to LiveKit Phone numbers where callee audio is not being heard by the caller. This is currently only experienced for numbers issues by LiveKit, and no other SIP provider connectivity should be affected. Our team is actively investigating this.

An upstream carrier reverted a codec change that was causing one-way audio issues on calls to LiveKit-issued phone numbers. As of 21:10 UTC calls are connecting normally and we are no longer seeing the issue. The upstream codec change has impacted calls from T-Mobile and AT&T carriers, and no other SIP provider connectivity was affected. We are monitoring before marking this resolved.

This incident has been resolved. We will follow up with a detailed postmortem as soon as possible.

Since ~06:30 UTC, outbound calls routed through Twilio are failing at an elevated rate. Twilio is reporting degraded voice connectivity (https://status.twilio.com/), affecting North America, Latin America, Europe, and the Middle East & Africa.

We are currently investigating reports of our dashboard failing to load across multiple regions, beginning at 0700 UTC. Other services remain unaffected and the impact is limited to the dashboard.

We have identified the cause: a problematic node in US-East and are deploying a fix. Impact continues to be limited to the dashboard.

A fix has been deployed as of 1055 UTC, and the dashboard is now loading successfully. We are continuing to monitor before marking this resolved.

This incident has been resolved and dashboard load times returned to normal by 1055 UTC