Dagychu Enterprise
Reports from system data, job outputs and ClickHouse
Dagychu Enterprise turns execution state and structured job results into saved, filterable reports. PostgreSQL remains the authoritative state store; ClickHouse is an optional reporting and BI destination.
What this page covers
This is a native Dagychu Enterprise capability, not a generic recommendation to connect an external database manually.
It covers two reporting needs: operational reports over Dagychu system data and business-facing views built from the structured output of executed jobs.
Operational reports over Dagychu system data
Saved Table Reports can query registered Dagychu views without giving every BI consumer direct access to the transactional schema.
- Task runs and task operations, including status and timing context.
- Job execution history with inputs, output_json, errors, durations and retry information.
- Pipeline registry, validation and pipeline drill-down views.
- Scheduler jobs, runs and events; worker and operational state; Overview incidents where applicable.
Turn output_json into a report
A Dagychu job can produce structured JSON. The Job Output View Designer reads output_json for the chosen job, pipeline or task scope and discovers available JSON paths.
An operator selects the fields that should become report columns, then applies filters and sorting. The saved configuration uses the expanded job-output report source, so teams can inspect business results without copying values from individual job pages.
| Stage | Dagychu behavior |
|---|---|
| Job runs | The execution retains task/job identity, input_json, output_json, status, timing and error context. |
| Fields are discovered | The designer enumerates JSON paths found in output_json for the selected scope. |
| View is configured | The operator chooses columns, filters and sorting instead of changing application code. |
| Report is saved | The configuration can be reopened and linked through a shareable report URL. |
| Data is consumed | Export directly or synchronize the saved report to ClickHouse for BI queries. |
Exports and shareable report state
Dagychu Enterprise report pages preserve useful view state so an operator can reproduce the same dataset instead of describing a sequence of clicks.
- Export report data as CSV, JSON or XLSX.
- Export SQL for controlled downstream use.
- Share a URL containing report sorting, filters and, where applicable, snapshot time.
- Retain readable filter conditions and report-time export headers.
Optional ClickHouse synchronization for BI
When ClickHouse is configured for Enterprise, saved Table Reports can be synchronized into the Dagychu reporting database. A read-only BI account can query those synchronized report tables without receiving write access to Dagychu.
If ClickHouse is not configured, report definitions and report use remain PostgreSQL-backed; synchronization is skipped. ClickHouse does not replace PostgreSQL as the authoritative Dagychu state store.
| Component | Responsibility |
|---|---|
| PostgreSQL | Authoritative product state, execution metadata and saved report configuration. |
| ClickHouse | Optional Enterprise destination for synchronized report tables and execution analytics. |
| Worker log storage | Full job logs consumed by the control plane; not replaced by ClickHouse report synchronization. |
| BI client | Read-only access to the reporting database when the operator enables and exposes that integration. |
Execution analytics in ClickHouse
The Enterprise execution-facts layer records operational facts such as pipeline and job identity, status, timestamps, duration and retry count. These facts support aggregate analysis such as failure hotspots and duration drift without turning the transactional state tables into a general analytics workload.
Security and operating boundary
Job outputs can contain sensitive business data. Scope saved reports deliberately, limit who can create or view them through Enterprise permissions, and expose any public BI URL only behind the controls appropriate for your environment.
- Use the provisioned read-only BI identity for consumption, not a write-capable ClickHouse account.
- Treat selected output_json fields as an explicit data contract and avoid exporting secrets.
- Review retention, network exposure and report scopes before connecting external BI tools.
Frequently asked questions
Does Dagychu have native ClickHouse reporting?
Yes, in Enterprise. Saved Table Reports can optionally synchronize to ClickHouse, and Enterprise execution facts support operational analytics.
Can I report on job output?
Yes. The Job Output View Designer discovers JSON paths in output_json, lets an operator choose report columns, filters and sorting, and saves the resulting report configuration.
Which export formats are available?
Dagychu Enterprise Table Reports support CSV, JSON, XLSX and SQL export.
Is ClickHouse the Dagychu state database?
No. PostgreSQL is the authoritative state store. ClickHouse is optional and used for synchronized reports and execution analytics in Enterprise.
Is this available in Community?
No. Community includes PostgreSQL-backed execution and operational views, but ClickHouse Table Reports and the saved BI reporting layer are Enterprise capabilities.
Evaluate Dagychu
Start with one real Python or Bash workflow. Verify a successful run, inspect a controlled failure, and test the recovery boundary against your own business result.