Supercharge Time-Series on Databricks
A toolkit that brings time-series capabilities to the Databricks Data Intelligence Platform — built on Lakebase, native to Unity Catalog.

Hot Lakebase, cold Unity Catalog — one query surface
Recent rows live in Lakebase for sub-10ms reads. Lakebase CDF streams older data into a Unity Catalog Managed Table for cheap, long-horizon retention.
-- One call turns any table into a time-partitioned ChronoTable CREATE TABLE metrics ( time TIMESTAMPTZ NOT NULL, device TEXT, cpu DOUBLE PRECISION ); SELECT lakets.create_chronotable( 'metrics', 'time', '1 day' ); -- Or: multi-metric table + chunks + indexes in one call SELECT lakets.create_metric_table( 'system_metrics', tag_columns := ARRAY['host', 'region', 'env'], field_columns := ARRAY['cpu', 'memory'], chunk_interval := '1 day' );
From regular table to ChronoTable in one call
LakeTS partitions your table by time, pre-creates future chunks, and adds BRIN indexes for fast time-range scans. The metadata registry wires the table into time_bucket, RollUps, and Lakehouse Sync automatically.
Drop a 30-day-old chunk in milliseconds — no row-by-row DELETE, no manual partition juggling.
Everything you need for time series at scale
ChronoTables
Time-partitioned tables with pre-created future chunks, BRIN indexes, and instant chunk drops for retention.
Learn more →Time series functions
time_bucket, first, last, locf, interpolate, delta, rate, gapfill — all native PL/pgSQL, no extensions.
Learn more →Incremental RollUps
Per-bucket DELETE+INSERT with watermark + invalidation log. DAG cascade refreshes hierarchical aggregates in topological order.
Learn more →Last Value Cache
Trigger-maintained "current state" table. Sub-10ms reads on the latest value per key, perfect for status widgets.
Learn more →Lakebase CDF
CDC replication to a Unity Catalog Managed Table via shadow tables. RollUps stay in Lakebase, raw data tiers out for long-horizon analytics.
Learn more →Alerts + Bulk Ingest
SQL-native threshold and deadman alerts. JSONB batch ingest from edge devices and Prometheus-compatible writers.
Learn more →