Update opensearch-release-notes-2.2.0.md (#2454)

Updates links in third release highlight

release-notes/opensearch-release-notes-2.2.0.md

@@ -3,7 +3,7 @@
 ## Release Highlights
 * OpenSearch 2.2.0 supports [logistic regression](https://github.com/opensearch-project/ml-commons/issues/318) and [RCFSummarize](https://github.com/opensearch-project/ml-commons/issues/356) machine learning algorithms.
 * With the [addition of the Lucene implementation](https://github.com/opensearch-project/k-NN/issues/380) of the HNSW algorithm, you can now choose from Lucene or the C-based Nmslib and Faiss libraries for approximate k-NN search.
-* You can now [search by relevance](https://github.com/opensearch-project/sql/issues/182) using SQL and PPL queries including [match_phrase_prefix](https://github.com/opensearch-project/sql/issues/186), [query_string](https://github.com/opensearch-project/sql/issues/191), and [highlights](https://github.com/opensearch-project/sql/issues/636).
+* You can now [search by relevance](https://github.com/opensearch-project/sql/issues/182) using SQL and PPL queries including [match_phrase_prefix](https://github.com/opensearch-project/sql/pull/661), [query_string](https://github.com/opensearch-project/sql/pull/675), and [highlight](https://github.com/opensearch-project/sql/pull/717).
 * You can now [upload your own custom region maps](https://github.com/opensearch-project/geospatial/issues/122) in GeoJSON format and use them for visualizations in OpenSearch Dashboards. You can also draw your own geographic boundaries on a visualization.
 * Several [rollup enhancements](https://github.com/opensearch-project/index-management/issues/408) allow you to roll up aggregated results from older data to dynamic target indexes and run one search query across multiple indexes.
 * You can now view [feature attribution and expected value](https://github.com/opensearch-project/anomaly-detection/issues/299) on the anomaly detection details page.