# Databricks — How Databricks Feature Store serves features with sub-second freshness

- Company: Databricks (databricks.com)
- Announced: 2026-08-17T23:10:00+00:00
- Category: not stated
- Coverage: not counted
- Announcement: yes
- Group: announcements
- Source: https://www.databricks.com/blog/how-databricks-feature-store-serves-features-sub-second-freshness
- Record: https://forck.live/items/17500-how-databricks-feature-store-serves-features-with-sub-second-freshness
- Subject: Mosaic AI

Databricks Feature Store brings real-time freshness to ML features: streaming aggregations from Kafka can now reach the online feature store with 200ms p99 latency, collapsing feature lag from minutes or hours to milliseconds. Spark Real-Time Mode (RTM) makes millisecond feature computation possible: RTM processes rows continuously instead of waiting for microbatches, updates rolling-window aggregates per event, and amortizes checkpointing to keep stateful streaming latency low. Lakebase enables high-throughput online feature writes: the separation of compute and storage layers reduces write amplification for frequent small upserts, making fresh feature values quickly available for low-latency model inference. Machine learning models are only as good as the signals they receive. A fraud detection use case must decide in milliseconds within a user pressing purchase whether to allow the transaction. Making the right call depends on seeing a suspicious transaction happening only seconds ago. Combining a user’s average transactions for the last 30 days along with the total transaction amount from the last 10 minutes highlights the potential fraud. …

---

Record: https://forck.live/items/17500-how-databricks-feature-store-serves-features-with-sub-second-freshness
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md