AWS AmazonCloudWatch documentation change

Service: AmazonCloudWatch · 2026-07-01 · Documentation low

File: AmazonCloudWatch/latest/monitoring/processor-examples.md

Summary

Added new 'Metrics pipeline examples' section with three examples: adding business context tags, removing high-cardinality attributes, and standardizing naming conventions.

Security assessment

The changes demonstrate general metric processing capabilities without addressing specific vulnerabilities or security incidents. The 'Remove high-cardinality attributes' example focuses on cost optimization rather than security remediation.

Diff

diff --git a/AmazonCloudWatch/latest/monitoring/processor-examples.md b/AmazonCloudWatch/latest/monitoring/processor-examples.md
index c1287a323..743bb4feb 100644
--- a//AmazonCloudWatch/latest/monitoring/processor-examples.md
+++ b//AmazonCloudWatch/latest/monitoring/processor-examples.md
@@ -6,0 +7,2 @@
+Metrics pipeline examples
+
@@ -9 +11,3 @@
-Here are common scenarios and example configurations for combining processors:
+Here are common scenarios and example configurations for combining processors.
+
+**Logs pipeline examples**
@@ -100,0 +105,43 @@ Remove sensitive fields only when the environment is production:
+## Metrics pipeline examples
+
+The following examples show processor configurations for metrics pipelines. Metrics processors use OTTL path expressions to target attributes at different scopes.
+
+###### Example Add business context to metrics
+
+Add team ownership and environment tags to metric datapoints:
+    
+    
+    processor:
+      - add_attributes:
+          attributes:
+            - key: resource.attributes["team"]
+              value: "platform-engineering"
+            - key: resource.attributes["cost_center"]
+              value: "CC-1234"
+
+###### Example Remove high-cardinality attributes
+
+Strip attributes that drive up storage costs. Does not apply to cumulative metrics or vended metrics — if any metrics in the selection criteria have unsupported temporality, the pipeline emits an `UnsupportedTemporality` warning metric that you can monitor in the `AWS/Observability Admin` namespace:
+    
+    
+    processor:
+      - delete_attributes:
+          with_keys:
+            - resource.attributes["host.id"]
+            - datapoint.attributes["http.request.id"]
+
+###### Example Standardize naming conventions
+
+Rename metrics and attributes to align with OpenTelemetry semantic conventions. Does not apply to cumulative metrics or vended metrics:
+    
+    
+    processor:
+      - rename_metrics:
+          metrics:
+            - from: "cpu_usage_percent"
+              to: "system.cpu.utilization"
+      - rename_attributes:
+          attributes:
+            - from_key: resource.attributes["hostname"]
+              to_key: resource.attributes["host.name"]
+