Monitoring Over Time
Page Outline
Returns a detailed timeseries breakdown of all monitored model health indicators, supporting fine-grained trend analysis and SLA tracking over custom time ranges.
Base Path: /api/v1/models
Method: GET
Path: /{model_id}/monitoring/over-time
Auth: Requires a valid Bearer token (Authorization: Bearer <JWT>)
Path Parameters
| Parameter | Type | Description |
|---|---|---|
model_id | string | Unique identifier of the model |
Query Parameters
| Parameter | Type | Default | Description |
|---|---|---|---|
from | string (ISO 8601) | 30 days ago | Start of the time window |
to | string (ISO 8601) | now | End of the time window |
granularity | string | 1d | Interval size: 1h, 6h, 1d, 7d |
indicators | string | all | Comma-separated list of indicators to include: drift, data_quality, throughput, latency, alert_count |
version_id | string | — | Scope to a specific version. Defaults to all version segments |
Behavior
- Returns one data point per
granularityinterval for each requested indicator. - Time intervals with no inference traffic return
nullvalues rather than0, to distinguish between zero activity and truly no data. - Intervals spanning version promotions are split into separate segments and annotated with the
version_idactive during that interval. - Returns
HTTP 404if the model is not found or monitoring has not been initialized.
Response
Returns a MonitoringTimeSeries object.
{
"model_id": "mdl-7f3a19bc",
"from": "2026-03-06T00:00:00Z",
"to": "2026-03-13T00:00:00Z",
"granularity": "1d",
"data_points": [
{
"timestamp": "2026-03-07T00:00:00Z",
"version_id": "ver-001",
"drift_score": 0.04,
"data_quality_rate": 0.989,
"throughput": 18420,
"avg_latency_ms": 17.2,
"alert_count": 0
},
{
"timestamp": "2026-03-12T00:00:00Z",
"version_id": "ver-001",
"drift_score": 0.24,
"data_quality_rate": 0.962,
"throughput": 21050,
"avg_latency_ms": 19.8,
"alert_count": 1
}
]
}