Investigate
When something looks wrong, find out why.
Open an investigation and Juno drives it with a dedicated toolkit. It pulls run history, audit logs, contract violations and lineage, runs the statistics, and hands back structured findings, cited evidence, a timeline and a causal diagram.
Lenses
Look at the problem from the right angle.
Every investigation starts with a lens, and you can blend several.
An example investigation
“Why did revenue drop on Tuesday?”
Juno traces the dashboard back to its source, finds where the chain broke, and lays out the evidence in order.
Timeline
Credentials rotated on orders_db, per the audit log.
First failed sync. The hourly import returns 401.
Third failure. A failure streak begins.
Freshness contract warns. Orders hasn’t updated in 6 hours.
Dashboard viewed. Revenue appears 18% lower than Monday.
Example data. Finding: update the connection credentials, then re-run the import.
Toolkits
The statistics and the checks, built in.
Statistical tools
Period comparisons, outlier detection, change points, correlation, error clustering, failure streaks, reliability stats, forecasting and periodicity.
Data-quality tools
Column profiling, duplicate and invalid-cell detection, freshness checks, recent-change review, connection health, blast radius, orphaned resources and contract coverage.
Findings
Share what you found.
Every investigation ends in structured findings with cited evidence, a timeline and a causal diagram. Save your approach as a reusable template, export the report, or share read-only findings with a link.
- Export to Markdown, plus PDF from Beam
- Templates and shareable findings from Ray
- Scheduled, recurring investigations from Aurora
Rotated credentials broke the orders import3 sourcesNo alert fired for the failure streak2 sourcesUpdate credentials, enable connection alertsowner: Ops