Repository navigation
Lakemeter OSS v0.1.1
Patch release introducing safer upgrades and correcting AI Parse estimate persistence. No database schema or data migration is required.
New capabilities
Version-aware upgrades
- Added
scripts/upgrade.shwithstatus,plan,doctor,apply, androllbackcommands. - Added installation discovery, release planning, prerequisite checks, and Semantic Versioning policy enforcement.
- Added immutable runtime staging, authenticated health verification, concurrency locks, resumable execution, automatic recovery, and manual rollback.
- Added Lakebase backup branches for future minor data updates and major schema migrations. Patch upgrades never modify Lakebase.
Before upgrading, update your repository checkout or download this release, then run ./scripts/upgrade.sh plan and ./scripts/upgrade.sh doctor.
Direct Lakebase Autoscaling provisioning
- New installations create Lakebase Autoscaling projects, branches, and endpoints directly.
- The installer and application use Autoscaling endpoint credentials and resource bindings.
Bug fixes
Serverless upgrader authentication
- Fixed
upgrade.sh applyverification from Databricks serverless jobs by exchanging the notebook credential for an OAuth token scoped to the target Databricks App. - Preserved authenticated health, database connectivity, and exact-version checks without the prior HTTP 401 rollback.
AI Parse estimate persistence
- Fixed AI Parse estimates so calculation method, complexity, DBU quantity, page count, mode, and page volume persist and clone correctly.
- Mapped API-facing AI Parse fields consistently to the existing database columns, preventing saved estimates from losing their calculated configuration.
Documentation updates
- Added a complete Upgrade Guide.
- Updated the installer and deployment documentation for direct Lakebase Autoscaling provisioning and the new upgrade workflow.
Verification
- Release policy and runtime checksums verified.
- 43 upgrade and schema-alignment tests passed.
- Documentation and frontend production builds passed.
- The complete
v0.1.0install, AI Parse reproduction, standard serverlessupgrade.sh apply, andv0.1.1persistence regression workflow passed in thelakemeterworkspace.