Skip to content
Syntasa

Product · Lakehouse Performance

Lakehouse tables that stay fast and tidy.

Bring existing Delta tables under management, optimize only what changed, keep files compact and well ordered, and answer eligible queries from metadata. All on open-source Delta Lake.

Talk with our teamSee Lens
Table health · sample data
Many small files → compact, well-organized files

Illustration: table health rises from 62 to 99 as many small files are compacted into a few well-organized ones.

The lakehouse suite

Manage, optimize, accelerate, and see the health.

  • Adoption

    Existing Delta tables come under management partition by partition. Nothing is rewritten up front.

  • Delta Optimize

    Compaction, Z-ORDER and liquid clustering in planned, resumable, incremental runs.

  • Metadata Accelerator

    Eligible queries answered from table metadata instead of scanning data files.

  • Delta Track

    Table health, trends, runs and small-file counts.

Delta Track

Table health you can see.

Trends, optimize runs and a dataset view, so you know which tables need care and whether it helped.

Dataset healthEvery managed table scored, with file shape and layout.

Delta table adoption

Start with the tables you already have.

Point Syntasa at existing Delta tables and they come under management partition by partition. Nothing is rewritten up front; each partition is optimized when it next needs it.

Adoption · partition by partition · sample data
  • Not yet managed
  • Under management

Illustration: existing partitions come under management one at a time; nothing is rewritten up front.

Incremental optimize · only what changed · sample data
  • Optimized
  • Changed since last run
  • Re-optimizing

Illustration: each run re-optimizes only the few partitions that changed since the last run.

Incremental optimize

Only what changed gets touched.

Each run looks at the partitions that changed since the last one and re-optimizes only those. Large tables stay tidy without re-processing history.

  • Planned runs, not surprises
  • Resumable if a run is interrupted
  • Works the same in every deployment model

Compaction · Z-ORDER · liquid clustering

Scattered small files in. Compact, ordered files out.

Three ways to organize a table, each shown before and after. The table health score rises as files are compacted and ordered.

  • Compaction

    Many small files become a few right-sized ones.

    Table health 58 before, 96 after.

  • Z-ORDER

    Rows sorted across files by the columns you filter on, so queries skip more files.

    Table health 61 before, 94 after.

  • Liquid clustering

    Data grouped into clusters that adapt as the table grows.

    Table health 57 before, 97 after.

Answering a query · sample data
Scan the dataReads every file
Answer from metadataReads far less

Metadata Accelerator

Some questions don't need a scan.

Many common queries can be answered from what the table already records about itself. The Metadata Accelerator answers eligible queries from metadata without scanning data files, so they come back faster and read far less.

In Lens

The benefit, on the record.

Lens shows time saved (modeled) and files compacted next to every other platform figure, so table care is visible rather than assumed.

  • Fewer small files and faster queries reduce compute
  • Tables stay standard Delta tables
  • No proprietary format, no lock-in
  • Time saved (modeled)≈ 41h 18mthis week, shown in Lens · sample data
  • Files compacted62k → 4.2klast 14 days · sample data

Performance you can see, anywhere you run.

Check the health of your tables.

We will look at a few of your Delta tables together and show what Delta Track and the optimizer would do.

Talk with our teamExplore the platformA working session with our engineers, on your data and constraints.