Operating system for Kastor Data

Schema changes are
deployments, not
prayers.

The control plane for Apache Iceberg that brings release discipline to your data infrastructure. Draft, review blast radius, publish with confidence.

The problem
Data teams ship schema changes the way software teams shipped code in 2005. No staging. No impact analysis. No rollback plan. A dropped column at 2pm becomes a Slack fire drill at 3pm.

The Kastor workflow

01

Draft on a branch

Isolate schema changes in a branch. Test modifications without touching production. Collaborate with your team before anything goes live.

02

Review the impact pack

See the blast radius before you ship. Every downstream consumer, every breaking compatibility risk, every estimated backfill cost. All in one view.

03

Publish with a rollback tag

Ship changes with an explicit rollback point. If something goes wrong, recover in seconds. No Slack archaeology, no guesswork.

The impact pack

Know exactly what ships before it ships.

■

Blast radius

Ranked list of every downstream consumer affected. See who breaks before they break. Drops, type changes, nullability shifts, partition transforms.

■

Cost estimation

Backfill costs calculated upfront. Know the compute and storage impact before you commit. No surprise bills at month-end.

■

Health scoring

File bloat detection, drift monitoring, and table health metrics. Catch degradation patterns before they cascade into production incidents.

Not another catalog

Traditional catalogs

✕ Track metadata, not change workflows
✕ No blast radius analysis before changes
✕ No cost estimation for backfills
✕ Rollback is an afterthought

Kastor

✓ Purpose-built for schema change management
✓ Full impact pack before every publish
✓ Compute and storage cost estimation
✓ Explicit rollback tags on every release

Data infrastructure deserves the same rigor as application code.

Kastor brings release discipline to the Iceberg ecosystem. No more ad-hoc changes. No more Slack fire drills.