fix(logs): optimize LogRecord memory by removing redundant context

CHANGELOG.md

@@ -12,6 +12,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## Unreleased
 
+- `opentelemetry-sdk`: Optimize `LogRecord` memory in `BatchLogRecordProcessor` by clearing the `Context` object before buffering.
+  ([#4957](https://github.com/open-telemetry/opentelemetry-python/issues/4957))
 - `opentelemetry-sdk`: Add file configuration support with YAML/JSON loading, environment variable substitution, and schema validation against the vendored OTel config JSON schema
   ([#4898](https://github.com/open-telemetry/opentelemetry-python/pull/4898))
 - Fix intermittent CI failures in `getting-started` and `tracecontext` jobs caused by GitHub git CDN SHA propagation lag by installing contrib packages from the already-checked-out local copy instead of a second git clone

opentelemetry-sdk/src/opentelemetry/sdk/_logs/_internal/export/__init__.py

@@ -14,6 +14,7 @@
 from __future__ import annotations
 
 import abc
+import copy
 import enum
 import logging
 import sys
@@ -286,8 +287,13 @@ def on_emit(self, log_record: ReadWriteLogRecord) -> None:
             if log_record.resource is not None
             else Resource.create({})
         )
+        # Shallow copy the API log record to break the reference to the potentially large context
+        # while keeping the original context intact for other processors.
+        api_log_record = copy.copy(log_record.log_record)
+        api_log_record.context = Context()
+
         readable_log_record = ReadableLogRecord(
-            log_record=log_record.log_record,
+            log_record=api_log_record,
             resource=resource,
             instrumentation_scope=log_record.instrumentation_scope,
             limits=log_record.limits,