AGE-237 : Adding VCS Metadata from .git (#61)

* Moving VCS info to resource attributes

* Added VCS Metadata via git SDK

Author: bhogayatakb

middleware/distro.py

@@ -228,9 +228,6 @@ def custom_record_exception(span: Span, exc: Exception):
     # Determine if the exception is escaping
     current_exc = sys.exc_info()[1]  # Get the currently active exception
     exception_escaped = current_exc is exc  # True if it's still propagating
-
-    mw_vcs_repository_url = os.getenv("MW_VCS_REPOSITORY_URL")
-    mw_vcs_commit_sha = os.getenv("MW_VCS_COMMIT_SHA")
 
     # Serialize stack info as JSON string since OpenTelemetry only supports string values
     stack_info_str = json.dumps(stack_info, indent=2)
@@ -244,8 +241,6 @@ def custom_record_exception(span: Span, exc: Exception):
             "exception.language": "python",
             "exception.stacktrace": traceback.format_exc(),
             "exception.escaped": exception_escaped,
-            "exception.vcs.commit_sha": mw_vcs_commit_sha or "",
-            "exception.vcs.repository_url": mw_vcs_repository_url or "",
             "exception.stack_details": stack_info_str,  # Attach full stacktrace details
         }
     )

middleware/resource.py

@@ -12,6 +12,8 @@
 from middleware.detectors.detector import process_detector_input, get_detectors
 from middleware.version import __version__
 import typing
+import os
+import git
 
 _logger = getLogger(__name__)
 
@@ -54,6 +56,23 @@ def create_resource(options: MWOptions):
                 "Skipped Custom attributes: parsing error expected format `abcd=1234,wxyz=5678`"
             )
 
+    mw_vcs_repository_url = os.getenv("MW_VCS_REPOSITORY_URL")
+    mw_vcs_commit_sha = os.getenv("MW_VCS_COMMIT_SHA")
+
+    # Fallback to git if env vars are missing (independent logic, single Repo call)
+    git_url, git_sha = None, None
+    if not mw_vcs_repository_url or not mw_vcs_commit_sha:
+        git_url, git_sha = get_git_info()
+    if not mw_vcs_repository_url and git_url:
+        mw_vcs_repository_url = git_url
+    if not mw_vcs_commit_sha and git_sha:
+        mw_vcs_commit_sha = git_sha
+
+    if mw_vcs_repository_url:
+        attributes["vcs.repository_url"] = mw_vcs_repository_url
+    if mw_vcs_commit_sha:
+        attributes["vcs.commit_sha"] = mw_vcs_commit_sha
+
     detectors: typing.List["ResourceDetector"] = [
         OTELResourceDetector(),
         ProcessResourceDetector(),
@@ -70,3 +89,19 @@ def create_resource(options: MWOptions):
         initial_resource=Resource.create(attributes),
     )
     return resources
+
+
+def get_git_info():
+    """
+    Returns (repository_url, commit_sha) from the local .git directory.
+    Returns (None, None) if not a git repo or info unavailable.
+    """
+    try:
+        repo = git.Repo(search_parent_directories=True)
+        commit_sha = repo.head.commit.hexsha
+        remote_url = None
+        if repo.remotes:
+            remote_url = repo.remotes.origin.url
+        return remote_url, commit_sha
+    except Exception:
+        return None, None

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "middleware-io"
-version = "2.2.1"
+version = "2.3.0"
 requires-python = ">=3.8"
 description = "Middleware's APM tool enables Python developers to effortlessly monitor their applications, gathering distributed tracing, metrics, logs, and profiling data for valuable insights and performance optimization."
 authors = [{ name = "middleware-dev" }]
@@ -40,6 +40,7 @@ dependencies =[
     "opentelemetry-resource-detector-azure==0.1.5",
     "opentelemetry-resourcedetector-gcp==1.7.0a0",
     "opentelemetry-resourcedetector-docker==0.4.0",
+    "gitpython>=3.1.0",
 
     "opentelemetry-instrumentation-aio-pika==0.48b0",
     "opentelemetry-instrumentation-aiohttp-client==0.48b0",