Repository navigation
Summary
This release marks a significant milestone in aligning the local development environment with production logic, refactoring IO Managers for better consistency, and hardening the codebase for Kubernetes deployments.
Key Changes
1. Local Development & IO Manager Alignment (v0.7.0 - v0.7.1)
- Centralized Database Configuration: Introduced
DatabaseConfigresource as the single source of truth for all IO Managers (S3, SQL, DuckDB, Filesystem). - DuckDB Concurrency Fix: Implemented exponential backoff retries in
DuckDBIOManagerto handle parallel write locks during local development. - Filesystem IO Manager Parity: Updated
FilesystemIOManagerto support "standalone" incremental loading, matching the S3 production logic. - Lazy Loading: Made
duckdba lazy import to prevent dependency failures in production environments (K8s) where it isn't installed.
2. Infrastructure & CI/CD Enhancements (v0.6.0 - v0.6.3)
- Resource Optimization: Increased memory limits and requests for Dagster pods to prevent OOM errors during heavy transformations.
- CI/CD Improvements: Added branch and SHA tags to Docker images for better traceability.
- K8s Stability: Resolved PostgreSQL authentication issues during schedule evaluation in Kubernetes.
3. Data Modeling & Integrity (v0.5.1 - v0.5.4)
- Schema Hardening: Updated
Tickermodels withBigIntegerfor timestamps and improved metadata-driven table configuration. - Type Safety: Resolved over 150+ linting and type-checking issues (Ruff, Mypy, Pyright).
- Documentation: Added comprehensive guides for DNS configuration and Git workflows.
Technical Debt & Fixes
- Standardized on
v*.*.*tagging for automated Docker builds. - Improved error handling in S3 loading logic and data merging.
- Refactored asset factories for better modularity.
Full Changelog: v0.5.0...v0.7.1