Wanted to double check if there is a way of retrieving the memory/disk usage for each collection via the python client.
I'm finding it hard, even via the API to get this info.
I was only able to retrieve it properly via the undocumented API at cluster level via the metrics endpoint, but this is only valid for 1 collection and I'm also expecting it to include overhead of the system itself -- so, not perfect.
curl 'https://cloud.qdrant.io/api/v1/accounts/<account id>/clusters/<cluster id>/metrics'
{
"nodes": [
{
"node_id": "qdrant-87c760c7-213d-420b-a3f1-c5ff4f48afd4-0",
"cpu": {
"avg": {
"current": "1.2859171573814976",
"1": "1.5038854089908655",
"3": "1.6912302836250612",
"24": "1.285450600137149"
},
"total": {
"value": "16.0",
"unit": "vcpu/core"
}
},
"ram": {
"avg": {
"current": "16492157009.92",
"1": "15784657135.458462",
"3": "15552328054.491428",
"24": "15175489617.92"
},
"total": {
"value": "137438953472",
"unit": "bytes"
}
},
"ram_rss": {
"avg": {
"current": "16422656081.92",
"1": "15750153691.766153",
"3": "15529118363.062857",
"24": "15171270737.92"
},
"total": {
"value": "137438953472",
"unit": "bytes"
}
},
"disk": {
"avg": {
"current": "852148224.0",
"1": "393330057.84615386",
"3": "243511881.14285713",
"24": "57344.0"
},
"total": {
"value": "549755813888",
"unit": "bytes"
}
}
},
{
"node_id": "qdrant-87c760c7-213d-420b-a3f1-c5ff4f48afd4-1",
"cpu": {
"avg": {
"current": "1.283758001276242",
"1": "1.3981456994494512",
"3": "1.4964670337601391",
"24": "1.2836337429098768"
},
"total": {
"value": "16.0",
"unit": "vcpu/core"
}
},
"ram": {
"avg": {
"current": "16317880401.92",
"1": "15703150516.381538",
"3": "15501643565.348572",
"24": "15175055441.92"
},
"total": {
"value": "137438953472",
"unit": "bytes"
}
},
"ram_rss": {
"avg": {
"current": "16253425745.92",
"1": "15670976436.381538",
"3": "15479801352.777143",
"24": "15170730065.92"
},
"total": {
"value": "137438953472",
"unit": "bytes"
}
},
"disk": {
"avg": {
"current": "782372864.0",
"1": "366799635.6923077",
"3": "234112731.42857143",
"24": "57344.0"
},
"total": {
"value": "549755813888",
"unit": "bytes"
}
}
},
{
"node_id": "qdrant-87c760c7-213d-420b-a3f1-c5ff4f48afd4-2",
"cpu": {
"avg": {
"current": "1.2841290958888705",
"1": "1.3490270654217091",
"3": "1.404827823006273",
"24": "1.2840706468315952"
},
"total": {
"value": "16.0",
"unit": "vcpu/core"
}
},
"ram": {
"avg": {
"current": "16417093713.92",
"1": "15748785312.68923",
"3": "15530089700.205715",
"24": "15174608977.92"
},
"total": {
"value": "137438953472",
"unit": "bytes"
}
},
"ram_rss": {
"avg": {
"current": "16357775441.92",
"1": "15719158629.612309",
"3": "15509924507.062857",
"24": "15170594897.92"
},
"total": {
"value": "137438953472",
"unit": "bytes"
}
},
"disk": {
"avg": {
"current": "737804288.0",
"1": "340555933.53846157",
"3": "210842185.14285713",
"24": "57344.0"
},
"total": {
"value": "549755813888",
"unit": "bytes"
}
}
}
]
}
Wanted to double check if there is a way of retrieving the memory/disk usage for each collection via the python client.
I'm finding it hard, even via the API to get this info.
I was only able to retrieve it properly via the undocumented API at cluster level via the metrics endpoint, but this is only valid for 1 collection and I'm also expecting it to include overhead of the system itself -- so, not perfect.
Currently I can get it via:
with the sample data bellow: