Searches Model Monitoring Stats generated within a given time window.
HTTP request
POST https://{service-endpoint}/v1beta1/{modelMonitor}:searchModelMonitoringStats
Where {service-endpoint}
is one of the supported service endpoints.
Path parameters
Parameters | |
---|---|
modelMonitor |
Required. ModelMonitor resource name. Format: |
Request body
The request body contains data with the following structure:
JSON representation |
---|
{ "statsFilter": { object ( |
Fields | |
---|---|
statsFilter |
Filter for search different stats. |
timeInterval |
The time interval for which results should be returned. |
pageSize |
The standard list page size. |
pageToken |
A page token received from a previous |
Response body
Response message for ModelMonitoringService.SearchModelMonitoringStats
.
If successful, the response body contains data with the following structure:
JSON representation |
---|
{
"monitoringStats": [
{
object ( |
Fields | |
---|---|
monitoringStats[] |
Stats retrieved for requested objectives. |
nextPageToken |
The page token that can be used by the next |
Authorization scopes
Requires the following OAuth scope:
https://www.googleapis.com/auth/cloud-platform
For more information, see the Authentication Overview.
IAM Permissions
Requires the following IAM permission on the modelMonitor
resource:
aiplatform.modelMonitors.searchModelMonitoringStats
For more information, see the IAM documentation.
SearchModelMonitoringStatsFilter
Filter for searching ModelMonitoringStats.
JSON representation |
---|
{ // Union field |
Fields | |
---|---|
Union field
|
|
tabularStatsFilter |
Tabular statistics filter. |
TabularStatsFilter
Tabular statistics filter.
JSON representation |
---|
{ "statsName": string, "objectiveType": string, "modelMonitoringJob": string, "modelMonitoringSchedule": string, "algorithm": string } |
Fields | |
---|---|
statsName |
If not specified, will return all the stats_names. |
objectiveType |
One of the supported monitoring objectives: |
modelMonitoringJob |
From a particular monitoring job. |
modelMonitoringSchedule |
From a particular monitoring schedule. |
algorithm |
Specify the algorithm type used for distance calculation, eg: jensen_shannon_divergence, l_infinity. |
ModelMonitoringStats
Represents the collection of statistics for a metric.
JSON representation |
---|
{ // Union field |
Fields | |
---|---|
Union field
|
|
tabularStats |
Generated tabular statistics. |
ModelMonitoringTabularStats
A collection of data points that describes the time-varying values of a tabular metric.
JSON representation |
---|
{
"statsName": string,
"objectiveType": string,
"dataPoints": [
{
object ( |
Fields | |
---|---|
statsName |
The stats name. |
objectiveType |
One of the supported monitoring objectives: |
dataPoints[] |
The data points of this time series. When listing time series, points are returned in reverse time order. |
ModelMonitoringStatsDataPoint
Represents a single statistics data point.
JSON representation |
---|
{ "currentStats": { object ( |
Fields | |
---|---|
currentStats |
Statistics from current dataset. |
baselineStats |
Statistics from baseline dataset. |
thresholdValue |
Threshold value. |
hasAnomaly |
Indicate if the statistics has anomaly. |
modelMonitoringJob |
Model monitoring job resource name. |
schedule |
Schedule resource name. |
createTime |
Statistics create time. A timestamp in RFC3339 UTC "Zulu" format, with nanosecond resolution and up to nine fractional digits. Examples: |
algorithm |
algorithm used to calculated the metrics, eg: jensen_shannon_divergence, l_infinity. |
TypedValue
Typed value of the statistics.
JSON representation |
---|
{ // Union field |
Fields | |
---|---|
Union field value . The typed value. value can be only one of the following: |
|
doubleValue |
Double. |
distributionValue |
Distribution. |
DistributionDataValue
Summary statistics for a population of values.
JSON representation |
---|
{ "distribution": value, "distributionDeviation": number } |
Fields | |
---|---|
distribution |
Predictive monitoring drift distribution in |
distributionDeviation |
Distribution distance deviation from the current dataset's statistics to baseline dataset's statistics. * For categorical feature, the distribution distance is calculated by L-inifinity norm or Jensen–Shannon divergence. * For numerical feature, the distribution distance is calculated by Jensen–Shannon divergence. |