Internalize opentelemetry-exporter-otlp-proto-common adapter code (#43)

.github/workflows/check-vendor.yml

@@ -0,0 +1,32 @@
+# This workflow will delete and regenerate the opentelemetry-exporter-otlp-proto-common code using scripts/vendor_otlp_proto_common.sh.
+# If generating the code produces any changes from what is currently checked in, the workflow will fail and prompt the user to regenerate the code.
+# For more information see: https://help.github.com/actions/language-and-framework-guides/using-python-with-github-actions
+
+name: Check OTLP Proto Common Vendored Code
+
+on:
+  push:
+    branches: [ "main" ]
+    paths:
+      - "scripts/vendor_otlp_proto_common.sh"
+      - "src/snowflake/telemetry/_internal/opentelemetry/exporter/**"
+      - ".github/workflows/check-vendor.yml"
+  pull_request:
+    branches: [ "main" ]
+    paths:
+      - "scripts/vendor_otlp_proto_common.sh"
+      - "src/snowflake/telemetry/_internal/opentelemetry/exporter/**"
+      - ".github/workflows/check-vendor.yml"
+
+jobs:
+  check-codegen:
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v3
+    - name: Run vendor script
+      run: |
+        rm -rf src/snowflake/telemetry/_internal/opentelemetry/exporter/
+        ./scripts/vendor_otlp_proto_common.sh
+    - name: Check for changes
+      run: |
+        git diff --exit-code || { echo "Code generation produced changes! Regenerate the code using ./scripts/vendor_otlp_proto_common.sh"; exit 1; }

CHANGELOG.md

@@ -3,6 +3,7 @@
 ## Unreleased
 
 * Upgrade OpenTelemetry Python dependencies to version 1.26.0
+* Vendored in adapter code from package opentelemetry-exporter-otlp-proto-common and replaced protobuf dependency with custom vanilla python serialization
 
 ## 0.5.0 (2024-07-23)
 

anaconda/meta.yaml

@@ -12,7 +12,6 @@ requirements:
     run:
         - python
         - opentelemetry-api ==1.26.0
-        - opentelemetry-exporter-otlp-proto-common ==1.26.0
         - opentelemetry-sdk ==1.26.0
 
 about:

benchmark/benchmark_serialize.py

@@ -0,0 +1,86 @@
+import google_benchmark as benchmark
+
+from util import get_logs_data, get_metrics_data, get_traces_data, get_logs_data_4MB
+
+from snowflake.telemetry._internal.opentelemetry.exporter.otlp.proto.common._log_encoder import encode_logs
+from snowflake.telemetry._internal.opentelemetry.exporter.otlp.proto.common.metrics_encoder import encode_metrics
+from snowflake.telemetry._internal.opentelemetry.exporter.otlp.proto.common.trace_encoder import encode_spans
+
+from opentelemetry.exporter.otlp.proto.common._log_encoder import encode_logs as pb2_encode_logs
+from opentelemetry.exporter.otlp.proto.common.metrics_encoder import encode_metrics as pb2_encode_metrics
+from opentelemetry.exporter.otlp.proto.common.trace_encoder import encode_spans as pb2_encode_spans
+
+"""
+------------------------------------------------------------------------------
+Benchmark                                    Time             CPU   Iterations
+------------------------------------------------------------------------------
+test_bm_serialize_logs_data_4MB      730591536 ns    730562298 ns            1
+test_bm_pb2_serialize_logs_data_4MB  702522039 ns    702490893 ns            1
+test_bm_serialize_logs_data             100882 ns       100878 ns         6930
+test_bm_pb2_serialize_logs_data          97112 ns        97109 ns         7195
+test_bm_serialize_metrics_data          114938 ns       114934 ns         6096
+test_bm_pb2_serialize_metrics_data      161849 ns       161845 ns         4324
+test_bm_serialize_traces_data           123977 ns       123973 ns         5633
+test_bm_pb2_serialize_traces_data       131016 ns       131011 ns         5314
+"""
+
+def sanity_check():
+    logs_data = get_logs_data()
+    metrics_data = get_metrics_data()
+    traces_data = get_traces_data()
+
+    assert encode_logs(logs_data).SerializeToString() == pb2_encode_logs(logs_data).SerializeToString()
+    assert encode_metrics(metrics_data).SerializeToString() == pb2_encode_metrics(metrics_data).SerializeToString()
+    assert encode_spans(traces_data).SerializeToString() == pb2_encode_spans(traces_data).SerializeToString()
+
+@benchmark.register
+def test_bm_serialize_logs_data_4MB(state):
+    logs_data = get_logs_data_4MB()
+    while state:
+        encode_logs(logs_data).SerializeToString()
+
+@benchmark.register
+def test_bm_pb2_serialize_logs_data_4MB(state):
+    logs_data = get_logs_data_4MB()
+    while state:
+        pb2_encode_logs(logs_data).SerializeToString()
+
+@benchmark.register
+def test_bm_serialize_logs_data(state):
+    logs_data = get_logs_data()
+    while state:
+        encode_logs(logs_data).SerializeToString()
+
+@benchmark.register
+def test_bm_pb2_serialize_logs_data(state):
+    logs_data = get_logs_data()
+    while state:
+        pb2_encode_logs(logs_data).SerializeToString()
+
+@benchmark.register
+def test_bm_serialize_metrics_data(state):
+    metrics_data = get_metrics_data()
+    while state:
+        encode_metrics(metrics_data).SerializeToString()
+
+@benchmark.register
+def test_bm_pb2_serialize_metrics_data(state):
+    metrics_data = get_metrics_data()
+    while state:
+        pb2_encode_metrics(metrics_data).SerializeToString()
+
+@benchmark.register
+def test_bm_serialize_traces_data(state):
+    traces_data = get_traces_data()
+    while state:
+        encode_spans(traces_data).SerializeToString()
+
+@benchmark.register
+def test_bm_pb2_serialize_traces_data(state):
+    traces_data = get_traces_data()
+    while state:
+        pb2_encode_spans(traces_data).SerializeToString()
+
+if __name__ == "__main__":
+    sanity_check()
+    benchmark.main()