← zurück zur Übersicht

2026-06-21T16-01-48Z_pinecone-S-filtered

pineconeStufe S · 10.00 GBfilteredStatus: ok

Metriken

Throughput
Latenz ⌀
23.99 ms
Latenz p50
21.00 ms
Latenz p95
31.00 ms
Latenz p99
58.00 ms
Recall@1
94.40%
Recall@10
94.56%
Recall@100
93.84%
Precision@10
94.56%
NDCG@10

Rohdaten

Konfiguration

Config-Name
pinecone-S-filtered
DB-Version
serverless
Image
Index-Typ
serverless
Index-Params
{"spec":"serverless","region":"aws-us-east-1","metric":"cosine"}
Dataset
S · 2.650.000 Vektoren · dim 1024 · Variante A · 10.00 GB
Queries
1.000 (Concurrency 1)

Lauf

Gestartet
21.06.2026, 18:01:48
Beendet
21.06.2026, 18:01:48
Dauer
0s
Index-Bauzeit
6.72s
Index-Größe
CPU ⌀ / Peak
— / — cores
RAM ⌀ / Peak
— / — MB
Disk read (Paging)
Disk write
Pod-RAM-Limit
Default
Warmup-Queries
K8s
Nodes
undefined

Notizen

measured: host-side (pinecone-bench)
latency_source: x-envoy-upstream-service-time (server-side)
client_latency_ms_p50: 501.4
response_duration_ms_p50: 116
n_vectors_actual: 2650000
gt_file: ground_truth_filter_rating_gte_4_ids.npy
gt_note: null
serverless: true
insert_time_s: 9372.5