Get AI Config metrics by variation
Retrieve usage metrics for an AI Config by config key, with results split by variation.
Path parameters
Query parameters
The starting time, as milliseconds since epoch (inclusive).
The ending time, as milliseconds since epoch (exclusive). May not be more than 100 days after from.
An environment key. Only metrics from this environment will be included.
Response
Metrics computed
Example response
[
{
"metrics": {
"judgeRelevance": 0.14894159,
"inputCost": 7.386281948385884,
"outputCost": 1.2315135367772556,
"generationSuccessCount": 5,
"inputTokens": 0,
"generationErrorCount": 2,
"thumbsDown": 9,
"judgeToxicity": 0.6846853,
"totalTokens": 1,
"judgeAccuracy": 0.10246457,
"generationCount": 5,
"timeToFirstTokenMs": 2,
"outputTokens": 6,
"thumbsUp": 7,
"durationMs": 3,
"satisfactionRating": 0.4145608
},
"variationKey": "variationKey"
}
]Changes
Changed in 1 of the 5 revisions of this API.15
- ●
deleted the
headerrequest parameterLD-API-Versionrequest-parameter-removed
- ○
api tag
AI Configsaddedapi-tag-added
- ○
api tag
AI Configs (beta)removedapi-tag-removed
- ○
added the optional property
items/metrics/judgeAccuracyto the response with the200statusresponse-optional-property-added
- ○
added the optional property
items/metrics/judgeRelevanceto the response with the200statusresponse-optional-property-added
- ○
added the optional property
items/metrics/judgeToxicityto the response with the200statusresponse-optional-property-added
This revision also has 13 changes that name no endpoint, such as unreferenced schemas being removed. See the revision's changelog
- ●