Faster PostgreSQL backfills
PostgreSQL backfills support up to 50 parallel workers, larger range-based chunks, and parallel backfills across partitions.
Read moreChangelog
- Performance
- Performance
MongoDB range backfills in the Dashboard
Configure per-collection MongoDB range backfills in Pipeline Settings, including chunk size and maximum parallelism.
Read more - Performance
Parallel UUID backfills for PostgreSQL
Eligible PostgreSQL tables with a single UUID primary key can use parallel range backfills split into roughly even shards.
Read more - Performance
MSSQL range backfill settings
MSSQL table Advanced settings let you enable range backfills and tune chunk size and parallelism for eligible integer-keyed tables.
Read more - Performance
Server-side filtering for MongoDB CDC
MongoDB database-wide change streams apply database and collection filters before events are decoded, cutting unnecessary change-stream traffic.
Read more - Performance
MongoDB network compression by default
MongoDB wire compression is on by default, preferring zstd and falling back to zlib when the connection URI names no compressor.
Read more - Performance
Higher-throughput Snowpipe Streaming
Snowpipe Streaming fans rows across stable channels and drains them concurrently, keeping per-channel retry and commit behavior.
Read more - Performance
MotherDuck multi-step merge
Multi-step merge optimization is available for MotherDuck as well as Snowflake through the destination advanced settings.
Read more - Performance
Oracle: index-driven backfills for tables without primary keys
When primaryKeysOverride points to a unique index, Oracle backfill queries include an index hint so the optimizer uses that index instead of a full scan.
Read more - Performance
Skip backfill per table
Tables can set skipBackfill: true to bypass the initial snapshot and go straight to streaming CDC, in the dashboard and the Terraform provider.
Read more