Use Case

Kafka → Delta table. Sub-5 second latency.

Stop maintaining a landing-zone-to-warehouse bridge. DataLynxr ingests Kafka, Kinesis, and Pulsar events directly into lakehouse tables — the same ones your analysts query.

Start Free Trial View Connectors

dlx-stream — ingest status

Stream: events.user-actions → s3://my-bucket/events
Format: Delta Lake · Exactly-once · ACID

✓ Batch #1042 committed  —  2.1s  — 4,880 events
✓ Batch #1043 committed  —  2.3s  — 5,102 events
✓ Batch #1044 committed  —  1.9s  — 4,991 events

Throughput:  48,230 events/s  (stable)
Lag:         0 messages  (consumer at head)
Latency P50: 2.1s  |  P99: 3.8s

Kafka, Kinesis, Pulsar

Native source connectors for all three major streaming platforms. No Kafka Connect plugin to manage, no Spark Structured Streaming job to maintain.

Exactly-once delivery

ACID transactions on the write path plus consumer offset management on the read path. Events are committed once — never duplicated, never dropped.

Sub-5 second latency

P99 end-to-end latency under 5 seconds from Kafka produce to Delta table read-availability. Tested at 50,000 events/second.

Schema registry integration

Avro and Protobuf schemas from Confluent Schema Registry or AWS Glue Schema Registry. Automatic schema evolution when new fields appear in the stream.

Same table analysts query

Streaming data lands in the same Delta/Iceberg table your SQL queries target. No separate real-time store, no sync job, no eventual consistency lag.

Replayable from any offset

Failed or delayed batches replay from the committed Kafka offset. Time-travel lets you query any historical state of the table to debug ingestion issues.

Retire your landing zone.

Connect a Kafka topic and start streaming to Delta tables in under 15 minutes.

Start Free Trial Read the Quickstart