QUA-1806: Clean up Enrichment Datastore FAQ and remove internal flag#1123
Conversation
Replace the sensitive `enrichment_only=true` reference and "write-back capabilities" jargon in the "Which connectors support enrichment?" answer with user-facing wording that points readers to the Supported Enrichment Datastores list. Remove em-dashes across linking, settings, unlinking, storage, operations, troubleshooting, and API sections; standardize `None` as bold None; tighten the Maximum Source Examples entry; reword the Scan flow bullet to mention records written to dedicated enrichment tables; align the prefix-conflict troubleshooting bullet with the existing prefix-uniqueness wording; and note per-container tables that remain after unlink under Append/Overwrite remediation.
Greptile SummaryThis PR cleans up the Enrichment Datastore FAQ page: it removes an exposed internal implementation flag (
Confidence Score: 4/5Documentation-only change that removes an internal implementation detail and standardises formatting; safe to merge with minor wording observations. The change is purely editorial — no code, no schema, no API contract is altered. The removal of the valid input range for "Maximum Source Examples per Anomaly" and the introduction of the undefined term "anomalous source container" are small wording gaps that could leave users without enough context, but neither causes incorrect information to be published. docs/source-datastore/enrichment-datastore/faq.md — the two small wording gaps noted in comments (undefined "container" term and removed range) are worth a quick second read. Important Files Changed
Flowchart%%{init: {'theme': 'neutral'}}%%
flowchart TD
A[Source Datastore] -->|Link| B[Enrichment Datastore]
A -->|Scan| C{Remediation Strategy}
C -->|None| D[Anomaly metadata only stored in Qualytics]
C -->|Append| E[Source records appended to dedicated enrichment tables]
C -->|Overwrite| F[Enrichment tables replaced with latest anomalous records]
E --> G[Per-container tables remain after unlinking]
F --> G
A -->|Unlink| H[Connection severed, data preserved in enrichment datastore]
H --> I[Manual cleanup required via database tools]
Reviews (1): Last reviewed commit: "docs(enrichment-faq): remove internal fl..." | Re-trigger Greptile |
| ### How do I clean up enrichment tables after unlinking? | ||
|
|
||
| Qualytics does not automatically delete enrichment tables when you unlink. If you want to remove the historical data, you must **manually drop** the enrichment tables (e.g., `_prefix_check_metrics`, `_prefix_failed_checks`, `_prefix_source_records`, `_prefix_scan_operations`) directly in the enrichment datastore using your database tools. | ||
| Qualytics does not automatically delete enrichment tables when you unlink. If you want to remove the historical data, you must **manually drop** the enrichment tables (e.g., `_prefix_check_metrics`, `_prefix_failed_checks`, `_prefix_source_records`, `_prefix_scan_operations`) directly in the enrichment datastore using your database tools. If you used Append or Overwrite remediation, Qualytics also creates one table per anomalous source container; those tables remain as well. |
There was a problem hiding this comment.
Undefined term "anomalous source container"
The new sentence introduces "one table per anomalous source container" without defining what "container" means in this context. Elsewhere in the doc the vocabulary is "tables", "schema", or "source datastore". A user who hasn't read the enrichment overview may not know whether "container" means a database table, a schema, a file path, or something else. Consider replacing it with the same terminology used in the rest of the page (e.g., "one table per source table that contained anomalous records") or linking to the glossary entry for "container".
| ### What is the "Maximum Source Examples per Anomaly"? | ||
|
|
||
| It controls how many actual source data rows are written to the enrichment datastore when a quality check fails. Default is 10. Range: 1 to 1,000,000,000. For practical recommendations on which values to use, see the [Source Examples: Practical Recommendations](introduction.md#source-examples-practical-recommendations){:target="_blank"} section. | ||
| It controls how many actual source data rows are written to the enrichment datastore when a quality check fails. Default is 10. For recommended values, see the [Source Examples: Practical Recommendations](introduction.md#source-examples-practical-recommendations){:target="_blank"} section. |
There was a problem hiding this comment.
Range information removed without replacement
The previous answer included "Range: 1 to 1,000,000,000" for Maximum Source Examples per Anomaly, which was actionable input-validation guidance. The new text only says "Default is 10" and redirects to a recommendations section. If the recommendations page does not restate the valid range, users who need to set a specific large value have no way to confirm the upper bound without consulting the product UI or API docs directly.
…r to glossary Address Greptile review on PR #1123: put the valid 1 to 1,000,000,000 range back on the Maximum Source Examples per Anomaly answer (the recommendations table covers values but not the explicit bounds), and link "container" in the post-unlink cleanup note to the glossary entry instead of redefining it inline.
Overview
Refines the Datastore Enrichment FAQ. Engineering flagged an internal flag exposed in the "Which connectors support enrichment?" answer; this PR removes that detail and aligns the rest of the page with our writing standards (no em-dashes, plain-language wording, consistent bold for UI values).
Key Changes
Noneas None in remediation references.Pages to Test