Skip to content

Changelog

Version history and release notes for Dango.

Version Numbering

Dango follows Semantic Versioning:

  • MAJOR.MINOR.PATCH (e.g., 1.0.0)
  • MAJOR: Breaking changes
  • MINOR: New features (backwards compatible)
  • PATCH: Bug fixes

Current Version

v1.0.7

Released: August 27, 2026

Status: Stable

Added

  • dango doctor — credential health check across all configured sources; shows ✓ OK / ✗ Missing / Expired / Unknown at a glance
  • Model wizard — interactive upstream table selection with CTE scaffold auto-generation; each selected table becomes a named CTE with {{ ref(...) }}
  • Schedule wizard — weekday preset (Mon–Fri pre-selected) in dango schedule add; dango schedule reload applies config changes without restart
  • Service account auth — dango oauth google_sheets and dango oauth google_analytics offer a choice between browser OAuth and JSON key file, useful for server deployments
  • Script UI history — scheduled script runs now appear in the Scripts page with last run timestamp, status, and View Log link (previously always showed "not run")
  • Script failure history — cancelled, timed-out, and pre-launch-failed script runs write history entries so failures are visible in the UI
  • Secrets page partial masking — env var values show the last 4 characters (****efgh) instead of blanket ***
  • Metabase Collection hierarchy — nested collections preserved correctly across dango metabase save / dango metabase load roundtrips
  • Column descriptions — default descriptions for columns in 7 high-value sources (Google Ads, GA4, Google Sheets, Stripe, HubSpot, Facebook Ads, BigQuery) visible in the catalog
  • Notebook role access — editor role can open notebooks not created by them
  • Scripts cloud sync — scripts and scripts/requirements.txt synced to remote server on dango remote push; invalid script paths rejected with a clear error

Fixed

  • Scripts page: scheduled runs no longer permanently show "not run" — history file written after every scheduled execution
  • Scripts page: View Log no longer 404s on scheduled runs — log files written to the correct path
  • Secrets page: duplicate /settings/variables page removed; env vars consolidated into /settings/secrets; old URL redirects (301) automatically
  • DuckDB lock conflicts on local: syncs now retry up to 5 times with 10s backoff when Metabase holds the DuckDB connection (previously failed immediately)
  • Schedule reload: source-list changes now detected and applied; previously dango schedule reload only compared cron triggers
  • Google Sheets: empty range now raises a clear error instead of silently bypassing empty-replace protection
  • Service account validation: no longer returns valid when google-auth package is not installed
  • Model wizard: duplicate CTE alias no longer generated when two upstream table names produce the same identifier
  • Sort/filter UI: broken on Schedules, Scripts, and Catalog pages after a JS refactor — now fixed
  • InquirerPy added to declared package dependencies — dango oauth google_sheets / dango oauth google_analytics no longer crash on fresh installs

Security

  • PostCSS updated 8.5.10 → 8.5.26, resolving path traversal vulnerabilities in the CSS build toolchain (build-time only; not exploitable at runtime)

v1.0.6

Released: August 19, 2026

Status: Stable — quality gate round

Added

  • Script scheduling — cron-based scheduling for Python scripts in scripts/ via dango schedule add
  • Scripts tab in web UI — list, run, cancel, and view logs for scripts
  • dango transform seed — runs dbt seeds; seed tables appear in catalog with row counts and profiling
  • Python 3.13 support
  • Schedule-aware staleness detection — sources with a schedule show a yellow "Stale" badge when last sync exceeds 2× the schedule interval
  • Per-table empty-replace protection for multi-resource dlt sources
  • Source and model row counts in web UI and catalog
  • Configurable backup retention, secrets exclusion from backups
  • Activity log deduplication and sync history gap prevention
  • dlt upgraded from 1.24.0 to 1.28.1

Fixed

  • Custom dlt_native sources in custom_sources/ now auto-discovered without manual config
  • Sources page crash when a dlt_native source has a null incremental capability
  • @dlt.source decorator crash during import inspection
  • "Incremental" sync mode label now derived from actual write_disposition in DuckDB
  • Models page: "Run" button no longer overwrites Schema column
  • Models page: "Last Run" column now sortable
  • Catalog: tab navigation no longer freezes when viewing tables filtered by source
  • Catalog: dbt docs 📖 link opens in new tab
  • DbtLock scope narrowed to write phase — long-running API syncs no longer block concurrent sources
  • DbtLock unlink race condition fixed
  • Metabase kept running during sync extract phase
  • Scheduled sync lock timeout aligned with direct sync (60s → 300s)
  • Notebook startup hang fixed; admin restart now shows warning modal
  • Scheduled sync "Unknown sync completed" toast fixed
  • Column alignment and filter inputs improved across Sources, Models, Schedules pages
  • DuckDB 1.5.2 → 1.5.4, Metabase JDBC driver 1.5.1.0 → 1.5.3.0, dbt-core 1.10.20 → 1.10.22

v1.0.5

Released: June 25, 2026

Added

  • Sync queue — concurrent syncs wait instead of failing, with queued status in the web UI
  • WebSocket sync phase events — UI shows "Processing..." during post-sync hooks
  • Structured logging to .dango/logs/dango.log with daily rotation and gzip compression
  • dango_version field in activity log, audit log, and sync subprocess headers
  • Schedules page: sources displayed as alphabetically sorted bullet list
  • OAuth timeout with 120s limit, retry option, and provider-specific troubleshooting

Fixed

  • OAuth credentials saved only after token exchange succeeds
  • Cross-project port kill — dango stop scoped to current project
  • Staging models generate explicit column lists instead of SELECT *
  • Sync log filenames include source name and timestamp
  • Notebook startup waits indefinitely for marimo to respond

v1.0.4

Released: June 15, 2026

Added

  • Global sync status indicator in UI header
  • Activity log entries for CSV uploads, deletes, and manual schedule triggers
  • "View in Metabase" link in catalog table detail
  • Empty sync protection — replace-mode syncs that return 0 rows preserve existing data
  • --allow-empty-replace CLI flag to override empty sync protection

Fixed

  • OAuth wizard shows success only after actual tokens obtained
  • Duplicate sync_started event emission removed

v1.0.3

Released: June 2026

Fixed

  • Metabase session bridging stability improvements
  • Schedule execution reliability fixes
  • OAuth refresh handling for long-running deployments

v1.0.2

Released: June 2026

Fixed

  • Cloud deployment stability fixes
  • Remote backup and restore improvements
  • Firewall and domain management fixes

v1.0.1

Released: June 2026

Fixed

  • Background Metabase schema sync on startup
  • Port conflict detection improvements
  • Initial sync reliability fixes

v1.0.0b4

Released: June 2026

Status: Beta 4 — bug fix round

See Upgrade Notes for b4 for breaking changes and upgrade steps.

Breaking Changes

  • GA4 column names: Type suffixes removed (sessions_integer → sessions, bounce_rate_float → bounce_rate, etc.). Update custom dbt models referencing old names.
  • Google Ads / GA4 data types: date columns now DATE (were VARCHAR/TIMESTAMPTZ), clicks/impressions now INTEGER (were VARCHAR). Requires --full-refresh per source.
  • GA4 default queries: events and conversions now include landingPage dimension.

Fixes

  • Scheduler now loads jobs from schedules.yml on startup
  • Staging YML files (sources_*.yml) no longer overwritten on sync
  • Cloud file sync now includes custom_sources/ and seeds/
  • Backups now include dbt models, custom sources, and .env
  • Metabase dashboard export captures all cards
  • Sync failures properly logged as errors with status in sync history
  • OAuth wizard error handling improvements
  • resolve_install_source() correctly detects PyPI installs

v1.0.0

Released: June 2026

Status: First major release

This is the v1.0.0 release — a complete rewrite and expansion of the Dango platform, adding authentication, cloud deployment, a web UI, scheduled syncs, data governance, and notebooks.

Highlights

  • 33 data sources (up from 8 wizard-supported in v0.1.0), with 25 in the interactive wizard
  • Authentication — password login, TOTP 2FA, API keys, three roles (admin/editor/viewer)
  • Cloud deployment — automated DigitalOcean provisioning and Bring Your Own Server (BYOS) support
  • Web UI — 16-page dashboard for monitoring syncs, health, data catalog, schedules, and secrets
  • Scheduled syncs — APScheduler-based with cron expressions and webhook notifications
  • Data governance — schema drift detection, PII scanning (Presidio), data catalog with column descriptions
  • Marimo notebooks — integrated reactive notebooks with DuckDB snapshot isolation
  • Monitoring — health metrics, sync history, capacity alerts
  • dango dev workflow — branch-based dbt development with isolated databases
  • Snapshots — SCD Type 2 change tracking via dbt snapshots, plus DuckDB point-in-time copies
  • Metabase SSO bridge — automatic login through the Web UI

What's New by Area

Authentication & Security (Phase 2)

  • Password authentication with bcrypt hashing
  • TOTP-based two-factor authentication
  • API key authentication for programmatic access
  • Three user roles: admin, editor, viewer (29 permissions)
  • Metabase SSO bridge — seamless login through Web UI
  • Audit logging for security events
  • Credential encryption for OAuth tokens

Cloud Deployment (Phase 3)

  • dango deploy — automated DigitalOcean droplet provisioning
  • BYOS (Bring Your Own Server) — deploy to any Ubuntu 22.04 server via SSH
  • Caddy reverse proxy with automatic HTTPS (Let's Encrypt)
  • Security hardening: fail2ban, SSH key-only, unattended-upgrades
  • dango remote push — sync config and dbt models to server
  • dango remote logs, dango remote status, dango remote history
  • Deploy journal (append-only JSONL) with git guardrails

Data Governance (Phase 4)

  • Schema drift detection — alerts when source schemas change
  • PII scanning via Microsoft Presidio (targeted entity types)
  • Data catalog — browsable table/column inventory
  • Column descriptions (user-editable via Web UI)
  • Monitor configuration via CLI and Web UI

Scheduling (Phase 5)

  • APScheduler-based sync scheduling with cron expressions
  • dango schedule add/remove/list/enable/disable
  • Webhook notifications on sync success, failure, or stale data
  • Configurable stale data thresholds
  • Misfire grace handling for missed schedules

Notebooks (Phase 6)

  • Marimo notebook integration
  • Built-in templates (EDA, time series, funnel analysis, cohort analysis)
  • DuckDB snapshot isolation — notebooks use read-only copies
  • Launch from Web UI or CLI (dango notebook open)

Web UI (Phase 7)

  • 16-page web dashboard at http://localhost:8800
  • Dashboard page — sync status overview, health summary
  • Sources page — source list, sync history, run syncs
  • Models page — dbt model status, run transformations
  • Health & Logs page — system health, disk usage, process status
  • Catalog page — data catalog browser with search
  • Monitoring page — schema drift, PII results, metrics
  • Schedules page — manage sync schedules
  • Notebooks page — launch and manage Marimo notebooks
  • Secrets & Admin page — credential management, user administration

CLI Additions

  • dango init — project initialization with admin password setup
  • dango deploy / dango remote push — cloud deployment
  • dango auth — user management (create, list, delete, reset-password)
  • dango oauth — OAuth provider management
  • dango schedule — sync scheduling
  • dango monitor — monitoring and governance
  • dango notebook — notebook management
  • dango dev — branch-based dbt development
  • dango snapshot — SCD Type 2 snapshots and DuckDB copies
  • dango upgrade — in-place version upgrades with migrations
  • dango schedule webhook — webhook notification management
  • dango config validate — configuration validation

Infrastructure

  • DuckDB version alignment checks (startup + pre-commit hook)
  • Process separation — syncs run in subprocesses to prevent lock contention
  • Single-worker uvicorn for cloud deployments (WebSocket compatibility)
  • Smoke test suite (100+ checks)

Breaking Changes from v0.1.0

  • dango init is now required before first use (sets up auth)
  • Authentication is enabled by default (use DANGO_ADMIN_PASSWORD env var for automation)
  • dango sync --source SOURCE syntax deprecated — use dango sync SOURCE (positional argument)
  • Metabase credentials now randomly generated and stored in .dango/metabase.yml
  • Project structure uses .dango/ directory for configuration (migrated automatically)

Previous Releases

v0.1.0

Released: December 17, 2025

Status: MVP Release

  • 8 wizard-supported data sources (CSV, Stripe, Google Sheets, GA4, Facebook Ads, Google Ads, REST API, dlt Native)
  • 25+ additional sources via dlt native configuration
  • Auto-generated dbt staging models
  • DuckDB warehouse with incremental loading
  • Metabase integration with auto-provisioned dashboards
  • Dashboard export/import (dango metabase save/load)
  • Web UI for monitoring
  • OAuth authentication for Google and Facebook sources

v0.0.5

Released: December 8, 2025

  • dango sync --dry-run to preview without executing
  • Unreferenced custom sources warning
  • Better validation output (database check, model count)

Upgrade Guide

Upgrading to v1.0.0

# Upgrade the package
dango upgrade

# Initialize auth (required for v1)
dango init

After upgrading:

  1. Set an admin password during dango init
  2. Review your sources — the config format is unchanged, but new features are available
  3. Access the Web UI at http://localhost:8800 after running dango start

General Upgrade

# Upgrade to latest
dango upgrade

# Or with pip directly
pip install --upgrade getdango

Release Process

How Releases Work

  1. Features developed on feature branches off main
  2. Merged to main after review
  3. Tagged releases published to PyPI
  4. Release notes added to this changelog

Release Cadence

  • Patch releases: As needed for bug fixes
  • Minor releases: As features complete
  • Major releases: When breaking changes are necessary

Post-v1 Roadmap

Planned for future releases:

  • Query performance logging and optimization insights
  • Health history — 24h/7d trend dashboards
  • Python task scheduler / reverse ETL
  • REST API provider presets (Shopify, Stripe templates)
  • OAuth callback unification
  • DuckLake integration (pending maturity evaluation)

Contributing

Want to contribute to the next release?

See the Contributing Guide for details.