diff --git a/ibm_db2/README.md b/ibm_db2/README.md
index 40ba7c8e4a6ad..fc2710cbd786c 100644
--- a/ibm_db2/README.md
+++ b/ibm_db2/README.md
@@ -24,7 +24,7 @@ The [ibm_db][4] client library is required. To install it, ensure you have a wor
 sudo -Hu dd-agent /opt/datadog-agent/embedded/bin/pip install ibm_db==3.2.6
 ```
 
-Note: If you are on an Agent running Python 2, use `ibm_db==3.0.1` instead of `ibm_db=3.1.0`.
+Note: If you are on an Agent running Python 2, use `ibm_db==3.0.1` instead of `ibm_db==3.2.6`.
 
 ##### Windows
 
diff --git a/ibm_db2/metadata.csv b/ibm_db2/metadata.csv
index 60fe5ce141dcd..27509d80f55a1 100644
--- a/ibm_db2/metadata.csv
+++ b/ibm_db2/metadata.csv
@@ -14,7 +14,7 @@ ibm_db2.bufferpool.group.column.hit_percent,gauge,,percent,,The percentage of ti
 ibm_db2.bufferpool.group.data.hit_percent,gauge,,percent,,The percentage of time that the group database manager did not need to load a page from disk to service a data page request.,1,ibm_db2,,
 ibm_db2.bufferpool.group.hit_percent,gauge,,percent,,The percentage of time that the group database manager did not need to load a page from disk to service a page request.,1,ibm_db2,,
 ibm_db2.bufferpool.group.index.hit_percent,gauge,,percent,,The percentage of time that the group database manager did not need to load a page from disk to service an index page request.,1,ibm_db2,,
-ibm_db2.bufferpool.group.xda.hit_percent,gauge,,percent,,The percentage of time that the group database manager did not need to load a page from disk to service an index page request.,1,ibm_db2,,
+ibm_db2.bufferpool.group.xda.hit_percent,gauge,,percent,,The percentage of XML storage object (XDA) pages that are found in memory (buffer pool) rather than requiring a physical read from disk.,1,ibm_db2,,
 ibm_db2.bufferpool.hit_percent,gauge,,percent,,The percentage of time that the database manager did not need to load a page from disk to service a page request.,1,ibm_db2,,
 ibm_db2.bufferpool.index.hit_percent,gauge,,percent,,The percentage of time that the database manager did not need to load a page from disk to service an index page request.,1,ibm_db2,,
 ibm_db2.bufferpool.index.reads.logical,count,,get,,"The number of index pages read from the logical table space containers for temporary, regular and large table spaces.",0,ibm_db2,,