10d • ❓ Q&A
run_ui crashes during LLM streaming
About 2 weeks ago I filed this bug in the Agent Zero repo https://github.com/agent0ai/agent-zero/issues/1658#issue-4506790033
Unfortunately I haven't got any response or comment to my bug report and I figured I would ask over here. Has anyone experienced anything like this running Agent Zero in a VM?
Basically the `run_ui` process inside the agent-zero container is abruptly crashing with exit status 247 and being restarted by the supervisor.
During the 2-3 seconds it takes to spawn back up, my cloudflared tunnel cannot route traffic.
Running docker inspect agent-zero explicitly confirms an OOM kill event: "OOMKilled":true.
Even though the container has an 8GB memory limit assigned by Docker, the Python process dies at ~1.9GB.
Are there internal cgroup or memory limits configured within the Agent Zero image/supervisor that I can override?
0
0 comments
run_ui crashes during LLM streaming
Leaderboard (30-day)
1
+7
2
+6
3
+4
4
+4
5
+4
Powered by