Managed-like UX for dpone¶
dpone includes self-service commands that help users diagnose their environment, generate manifests, preview execution plans, inspect state, write run artifacts, certify connectors, tune performance, and scaffold community connectors.
Fast path¶
dpone doctor --profile local
dpone init \
--source-type postgres \
--sink-type mssql \
--source-connection pg_src \
--sink-connection mssql_dwh \
--source-schema public \
--source-table orders \
--target-schema landing \
--target-table orders \
--strategy incremental_merge \
--unique-key id \
--out examples/batch/landing_orders.batch.yaml
dpone plan examples/batch/landing_orders.batch.yaml --selector public.orders
dpone perf advise examples/batch/landing_orders.batch.yaml --selector public.orders
Commands¶
dpone doctorchecks local readiness: Python, optional extras, Docker,bcp,sqlcmd, ClickHouse tools, Kafka/Schema Registry hints, and fix commands.dpone initgenerates a batch manifest bundle,.env.example, and smoke command.dpone planrenders a dry-run execution plan without changing targets.dpone run-reportwrites JSON, Markdown, and HTML run artifacts under.dpone/runsby default.dpone stateexposes safe inspect/reset/export/replay/compare UX. Destructive operations preview unless--yesis supplied.dpone connectorslists, certifies, and scaffolds connector SDK skeletons.dpone perf adviserecommends native bulk paths, partitioning, and delivery tuning.dpone studioprints local Studio API metadata by default;--servestarts the headless local HTTP API bridge for the separate Nuxt/Vuedpone-studioUI.
Manifest additions¶
All managed UX sections are optional and backward-compatible:
quality:
checks:
- type: min_rows
value: 1
mode: warn
observability:
artifacts:
enabled: true
formats: [json, md, html]
opentelemetry:
enabled: false
prometheus:
enabled: false
performance:
advisor:
enabled: true
certification:
connector_status: experimental
Safety defaults¶
dpone planis dry-run only.dpone state resetanddpone state replay-fromrequire--yesto execute; without it they return a preview payload.- Secrets are redacted in doctor, plan, run artifact, certification, and scaffold diagnostics.
- Studio is local-only in this version and has no SaaS backend.
Studio architecture¶
dpone intentionally does not embed a SaaS/product frontend. The framework repo owns the headless local API bridge, while the UI lives in a separate Nuxt/Vue repo:
- framework/API:
dpone studio --serve; - frontend repo:
PaulKov/dpone-studio; - contract: JSON endpoints listed below.
This keeps the ETL framework, connector runtime and frontend product independently deployable.
Studio API endpoints¶
Start the local Studio API bridge:
/returns a minimal API status page and points users to the separate Nuxt/Vue Studio repo./api/studioreturns Studio navigation, panels, quick actions and endpoint metadata for the frontend./api/doctorreturns redacted environment diagnostics./api/connectorsreturns local connector/feature metadata./api/perfreturns the same service-backed metadata envelope used by the performance panel.
Run artifacts¶
A run artifact contains timeline, state before/after, quality results, schema evolution notes, reconciliation metrics, and redacted diagnostics. Use JSON for automation, Markdown for repo evidence, and HTML for local review.
Connector SDK¶
The scaffold includes connector/source/sink skeletons, a contract test, and a docs template. Fill in auth, capabilities, examples, and certification evidence before publishing a community connector.
Production checklist¶
- Run
dpone doctor --profile productionon the execution host. - Run
dpone planand inspect staging/shadow/schema/state sections. - Run
dpone perf advisefor large tables or Kafka batches. - Configure
quality.checksfor correctness gates. - Store run artifacts for every release or live certification run.