i think i found it. we had to replace elasticsearch after install of clearml. then i guess clearml migrations iddn't rerun

version 7.14.4

this seems to be confirmed by this documentation None If you have not changed the default runtime on your GPU nodes, you must explicitly request the NVIDIA runtime by setting runtimeClassName: nvidia in the Pod spec:

i sniffed the traffic

on kubernetes

also the inability to see workers/queues has the same error but on diferent fields. so i guess i must be missing something bigger than just a misconfigured index ?

i did this as a workaround:

curl -XPUT " None " -H 'Content-Type: application/json' -d'
{
"properties": {
"metric": { "type": "text", "fielddata": true },
"variant": { "type": "text", "fielddata": true }
}
}'

but this workaround should not be needed ,right ? is this a compat issue ? or was my elasticsearch not properly initialized ?