- Promote stable, frequently filtered fields to typed columns.
- Keep long-tail and drifting attributes in VARIANT.
- Choose the event-time column, key and partition lifecycle deliberately.
- Add inverted indexes to message and proven retrieval fields—not every property.
All use cases
SUPPORTING CAPABILITY
Onboard unfamiliar logs with AI assistance
Turn a representative JSON event into a reviewable Doris schema and collector configuration without handing AI uncontrolled production access.
Discuss this use caseThe data exists; the pipeline does not
A new service emits useful events, but nobody wants to spend another sprint choosing types, indexes, retention and shipper syntax.
- Fields and types must be inferred manually.
- Schema drift makes rigid mappings fragile.
- Collector configuration differs by source.
- Unsafe automation creates more risk than it removes.
The UnifyLogs path
AI proposes. Your team decides.
- 01
Paste a sample
Provide a representative raw JSON log in the ingestion workbench.
- 02
Review the proposal
AI returns table DDL and a collector configuration in editable code blocks.
- 03
Deploy deliberately
An authorized user reviews and explicitly triggers table creation and a test event.
What changes
What changes for your team?
- Faster first ingestion for custom formats
- Reviewable output instead of opaque automation
- Flexible handling for evolving JSON attributes
Technical details: data flow and build plan
BUILD PLAN / ASSISTED ONBOARDING
Let AI remove the blank page—not the review gate
Give the workbench representative events, not credentials to production. Treat the proposed schema and config as an editable starting point and keep deployment explicit.
01Representative JSON
02Schema + config proposal
03Human review
04Deploy · test · observe
- Paste multiple representative samples into the ingestion workflow.
- Review generated DDL and collector config side by side.
- Edit types, retention, buckets and indexes before approval.
- An authorized operator explicitly creates the table and sends a test event.
BEFORE CUTOVER
Acceptance gates
- 1Samples cover optional and malformed fields
- 2DDL passes human review
- 3Collector config validates
- 4Test event lands with the expected types
LET’S TALK ABOUT YOUR LOGS
Where are your logs slowing you down?
Tell us about your current tools and the problem you want to solve. We’ll agree the next step: a product walkthrough or a technical evaluation with one source.