[processor/transform] Add ability to truncate attributes

CHANGELOG.md

@@ -17,6 +17,7 @@
 
 - `k8sclusterreceiver`: Validate that k8s API supports a resource before setting up a watcher for it (#9523)
 - `internal/stanza`: Add support for `remove` operator (#9524)
+- `transformprocessor`: Add new `truncation` function to allow truncating string values in maps such as `attributes` or `resource.attributes` (#9546)
 
 ### 🧰 Bug fixes 🧰
 
@@ -55,7 +56,7 @@
 - `datadogexporter`: Update deprecation messages to reflect new deprecation plan (#9422)
 - `cmd/mdatagen`: Update generated functions to have simple parse function to handle string parsing consistently and limit code duplication across receivers (#7574)
 - `attributesprocessor`: Support filter by severity (#9132)
-- `processor/transform`: Add transformation of logs (#9368)
+- `transformprocessor`: Add transformation of logs (#9368)
 - `datadogexporter`: Add `metrics::summaries::mode` to specify export mode for summaries (#8846)
 - `prometheusreceiver`: Add resource attributes for kubernetes resource discovery labels (#9416)
 

processor/transformprocessor/README.md

@@ -22,6 +22,8 @@ it references an unset map value, there will be no action.
 - `keep_keys(target, string...)` - `target` is a path expression to a map type field. The map will be mutated to only contain
 the fields specified by the list of strings. e.g., `keep_keys(attributes, "http.method")`, `keep_keys(attributes, "http.method", "http.route")`
 
+- `truncate_all(target, limit)` - `target` is a path expression to a map type field. `limit` is an integer.  The map will be mutated such that all string values are truncated to the limit. e.g., `truncate(attributes, 100)` will truncate all string values in `attributes` such that all string values have less than or equal to 100 characters.  Non-string values are ignored.
+
 Supported where operations:
 - `==` - matches telemetry where the values are equal to each other
 - `!=` - matches telemetry where the values are not equal to each other
@@ -48,6 +50,8 @@ processors:
         - set(status.code, 1) where attributes["http.path"] == "/health"
         - keep_keys(resource.attributes, "service.name", "service.namespace", "cloud.region")
         - set(name, attributes["http.route"])
+        - truncate_all(attributes, 4096)
+        - truncate_all(resource.attributes, 4096)
 service:
   pipelines:
     logs:
@@ -73,3 +77,5 @@ All spans
 1) Set status code to OK for all spans with a path `/health`
 2) Keep only `service.name`, `service.namespace`, `cloud.region` resource attributes
 3) Set `name` to the `http.route` attribute if it is set
+4) Truncate all span attributes such that no string value has more than 4096 characters.
+5) Truncate all resource attributes such that no string value has more than 4096 characters.