davidzollo opened a new issue, #11355:
URL: https://github.com/apache/seatunnel/issues/11355

   ## Background
   SeaTunnel has been improving CDC schema evolution correctness in several 
areas, including policy handling, multi-table routing, comment-related schema 
changes, and failure handling.
   
   That work is important, but users still lack an operator-facing way to 
answer a simple question:
   
   **What exactly happened to my schema change, and why did the pipeline make 
that decision?**
   
   Today, when a DDL-related problem occurs, users often have to reconstruct 
the story from logs, connector behavior, sink capability assumptions, and 
trial-and-error.
   
   ## Problem to solve
   SeaTunnel does not yet provide a first-class **schema evolution timeline and 
decision log**.
   
   For production CDC jobs, operators should be able to inspect:
   - when a schema change event arrived
   - how the event was resolved to a table identifier/path
   - which behavior policy applied
   - whether the sink claimed support for that change type
   - whether the schema apply succeeded, was ignored, or failed
   - whether the event was acknowledged consistently across the pipeline
   
   Without this, schema evolution remains hard to debug even when the 
underlying behavior is technically correct.
   
   ## Proposed scope
   Add a schema evolution timeline and decision log for CDC jobs.
   
   Suggested first-delivery scope:
   - record and present the main lifecycle of a schema change event from source 
ingestion to sink-side handling
   - show the event type, affected table, resolved target, policy decision, and 
final outcome
   - make failures and ignored events explicit instead of leaving them implicit 
in logs
   - support multi-table paths where routing mistakes or ambiguity can occur
   - connect the timeline to the job detail experience so users can inspect 
recent DDL activity without leaving the runtime context
   
   ## Why this needs explicit design first
   This feature crosses multiple contracts:
   - CDC source event normalization
   - table resolution / routing
   - schema-change behavior policy
   - sink capability declaration and apply behavior
   - user-facing API/UI representation of those decisions
   
   If SeaTunnel exposes this feature without a clear design, it risks showing 
incomplete or misleading state about schema evolution.
   
   ## STIP requirement before implementation
   Because this feature introduces a cross-module decision/audit model for CDC 
schema evolution, **the claimant should submit a STIP design first and get 
maintainer agreement before starting implementation**.
   
   The STIP should explicitly define:
   - the event lifecycle model shown to users
   - which fields are stable product contracts versus debug-only details
   - how policy decisions are represented
   - how sink support, apply success, ignore, retry, and failure states are 
modeled
   - retention boundaries and how much schema-event history is kept
   - how the design interacts with multi-table sinks and table-resolution edge 
cases
   
   ## Acceptance criteria
   - Users can inspect recent schema evolution activity as a timeline instead 
of inferring it from logs.
   - The timeline clearly shows event type, target table, policy decision, and 
final outcome.
   - Failures and ignored events are distinguishable.
   - The first version documents scope limits and retention behavior clearly.
   - English and Chinese docs are updated.
   
   ## Non-goals for the first version
   - full payload archival for every raw DDL statement forever
   - a guarantee that every connector exposes identical schema-event detail on 
day one
   - automatic schema repair suggestions in the first delivery
   
   ## Related work
   - Issue #11043
   - Issue #10776
   - Issue #11016
   - PR #11015
   - PR #11025
   - PR #11162
   - PR #11328
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to