Bulk ingest

Server-side batch ingest for high-throughput data loading from edge devices, microservices, or protocol adapters.

ingest_batch(p_table_name, p_data, p_schema_name)

Inserts multiple rows from a JSONB array. Designed for edge devices or microservices that batch measurements.

Parameter	Type	Default	Description
p_table_name	TEXT	—	Target ChronoTable
p_data	JSONB	—	Array of row objects
p_schema_name	TEXT	'public'	Schema

Returns: INT — number of rows inserted

SELECT lakets.ingest_batch('sensor_data', '[
    {"time": "2026-04-09T10:00:00Z", "device_id": "d1", "temperature": 22.5, "humidity": 60},
    {"time": "2026-04-09T10:00:01Z", "device_id": "d2", "temperature": 23.1, "humidity": 58},
    {"time": "2026-04-09T10:00:02Z", "device_id": "d3", "temperature": 21.8, "humidity": 62}
]');
-- Returns: 3

Bulk inserts trigger LakeTS's statement-level invalidation, so RollUps that depend on this ChronoTable have the affected buckets marked dirty automatically.

ingest_prometheus(p_table_name, p_metric_name, p_labels, p_value, p_timestamp, p_schema_name)

Inserts a single Prometheus-style metric. The target table must have columns: time, metric_name, labels (JSONB), value.

Parameter	Type	Default	Description
p_table_name	TEXT	—	Target table
p_metric_name	TEXT	—	Metric name
p_labels	JSONB	—	Label key-value pairs
p_value	DOUBLE PRECISION	—	Metric value
p_timestamp	TIMESTAMPTZ	now()	Observation timestamp
p_schema_name	TEXT	'public'	Schema

Returns: VOID

SELECT lakets.ingest_prometheus(
    'prom_metrics',
    'http_requests_total',
    '{"method": "GET", "path": "/api/v1/data", "status": "200"}'::JSONB,
    1542.0
);