hey team, working on a POC use-case. Currently in the experimenting phase. Our Livekit cloud hosted agent was working fine till now with no as such agent dispatch error, but recently we are facing this error a lot.
This error is faced both when I test it out in the Livekit Dashboard Console. And also in hosted playground where other team members are testing the agent (via a room token and wss URL)
Currently automatic dispatch is on, which we plan on removing as we move to production. Just wanted to know if this a plan issue as in do we need to upgrade or something else.
Really looking forward to your reply
Nothing immediately jumps out to me, most typically this is because you have exceeded your quotas but it looks like you have fewer than 5 simultaneous sessions (the maximum allowed under build). Did you make any recent changes to your agent that might cause it to crash when it is dispatched? The easiest way to determine the root cause is the check the agent logs when the agent is dispatched:
lk agent logs
More information:
@shashank.salian, Building on @darryncampbell’s lk agent logs pointer: “was working, now failing” patterns on Cloud Agents almost always reduce to one of two things visible in those logs.
- Silent entrypoint crash on a recent deploy. Worker registers fine (so quotas look OK), but the entrypoint raises on connect (missing env var, lazy-import error, plugin init failure). Look for a traceback at the top of the agent log around the time the error started.
- Prewarmed processes holding partial state. If you deployed without a clean restart, prewarm workers may be running mid-state. Force
lk agent deploy (or restart the deployment) to flush the pool.
If neither shows up, paste the actual error string from the screenshot here. That narrows it quickly.
