On the other side I have fleet running on kubernetes across several clusters for redundancy. Sometimes on-demand queries will work and sometimes I'll get the below error:
proxy_set_header Upgrade $http_upgrade; proxy_set_header Connection $connection_upgrade;
If I reduce the kubernetes deployments to a single deployment, everything works fine. Given that it's intermittent and works with a single cluster deployment, I suspect there is a handshake/routing issue happening at the kubernetes layer where responses are not landing where fleet expects them. Anyone got any thoughts on this?
Failed to load resource: the server responded with a status of 404 () bundle-ec4e3a98f99eeaa1006b.js:110 WebSocket connection to '<wss://fleetdm-testing.example.com/api/v1/fleet/results/198/dd3kiay2/websocket>' failed: