From c4866a959c37e9fa2abbd1dc360ca7c375898f64 Mon Sep 17 00:00:00 2001 From: Chris Collins Date: Thu, 3 Apr 2025 12:16:40 -0400 Subject: [PATCH] fix(docs) Update impact analysis docs to call out lightning cache bugs (#12918) --- docs/act-on-metadata/impact-analysis.md | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/docs/act-on-metadata/impact-analysis.md b/docs/act-on-metadata/impact-analysis.md index 3dbf532b2d..f1fc34801b 100644 --- a/docs/act-on-metadata/impact-analysis.md +++ b/docs/act-on-metadata/impact-analysis.md @@ -56,6 +56,14 @@ Follow these simple steps to understand the full dependency chain of your data e

+### Known Issues + +Impact Analysis is a powerful feature that can place significant demands on the system. To maintain high performance when handling large result sets, we've implemented "Lightning Cache" - an alternate processing path that delivers results more quickly. By default, this cache activates with simple queries when there are more than 300 assets in the result set. You can customize this threshold by setting the environment variable `CACHE_SEARCH_LINEAGE_LIGHTNING_THRESHOLD` in your GMS pod. + +However, the Lightning Cache has a limitation: it may include assets that are soft-deleted or no longer exist in the DataHub database. This occurs because lineage references may contain "ghost entities" (URNs without associated data). + +Note that when you download Impact Analysis results, our system properly filters out these soft-deleted and non-existent assets. As a result, you might notice differences between what appears in the UI and what appears in your downloaded results. + ## Additional Resources ### Videos