ci: Emit CloudWatch metrics from rust benchmarks (#4742)

.github/bin/criterion_to_cloudwatch.py

@@ -0,0 +1,126 @@
+# Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License").
+# You may not use this file except in compliance with the License.
+# A copy of the License is located at
+#
+#  http://aws.amazon.com/apache2.0
+#
+# or in the "license" file accompanying this file. This file is distributed
+# on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either
+# express or implied. See the License for the specific language governing
+# permissions and limitations under the License.
+#
+
+import argparse
+import json
+from typing import Tuple
+from collections.abc import Generator
+import boto3
+
+
+class CriterionResult:
+    def __init__(self, json_obj: dict):
+        self._json_obj: dict = json_obj
+
+    @property
+    def _group_and_category(self) -> Tuple[str, str]:
+        id_ = self._json_obj["id"]
+
+        # The id is made up of the bench group and the category, separated by a /.
+        # For example: handshake-ecdsa384/s2n-tls
+        id_split = id_.split("/")
+        assert len(id_split) == 2
+
+        group = id_split[0]
+        category = id_split[1]
+
+        return group, category
+
+    @property
+    def group(self) -> str:
+        group, category = self._group_and_category
+        return group
+
+    @property
+    def category(self) -> str:
+        group, category = self._group_and_category
+        return category
+
+    @property
+    def mean_us(self) -> float:
+        mean = self._json_obj["mean"]
+
+        unit = mean["unit"]
+        assert unit == "ns"
+
+        # CloudWatch doesn't support nanoseconds, so the units are converted to microseconds.
+        estimate_ns = mean["estimate"]
+        estimate_us = estimate_ns / 1_000
+
+        return estimate_us
+
+
+class CriterionReader:
+    def __init__(self, criterion_output_path: str):
+        self.criterion_output_path: str = criterion_output_path
+
+    def read_bench_results(self) -> Generator[CriterionResult, None, None]:
+        with open(self.criterion_output_path, 'r') as output:
+            for line in output.readlines():
+                obj = json.loads(line)
+
+                # Skip criterion output that isn't describing a benchmarking result.
+                if obj.get("reason") != "benchmark-complete":
+                    continue
+
+                yield CriterionResult(obj)
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser(
+        prog="criterion_to_cloudwatch",
+        description="Emits CloudWatch metrics from criterion output.",
+    )
+    parser.add_argument(
+        "--criterion_output_path",
+        required=True,
+        help="File containing criterion json output. Generated with cargo-criterion by specifying the "
+             "`--message-format json` option.",
+    )
+    parser.add_argument(
+        "--namespace",
+        required=True,
+        help="CloudWatch namespace to emit metrics to (e.g. s2n-tls-bench).",
+    )
+    parser.add_argument(
+        "--platform",
+        required=True,
+        help="Specifies the platform dimension of the CloudWatch metrics (e.g. Ubuntu24-x86).",
+    )
+    args = parser.parse_args()
+
+    cloudwatch = boto3.client("cloudwatch")
+
+    reader = CriterionReader(args.criterion_output_path)
+    for result in reader.read_bench_results():
+        print(f"Emitting {args.namespace} - {result.group}/{result.category} for {args.platform}: {result.mean_us}")
+
+        cloudwatch.put_metric_data(
+            Namespace=args.namespace,
+            MetricData=[{
+                "MetricName": result.group,
+                "Dimensions": [
+                    {
+                        "Name": "Category",
+                        "Value": result.category,
+                    },
+                    {
+                        "Name": "Platform",
+                        "Value": args.platform,
+                    },
+                ],
+                "Value": result.mean_us,
+                "Unit": "Microseconds",
+            }],
+        )

.github/workflows/bench.yml

@@ -0,0 +1,48 @@
+name: Benchmarking
+
+on:
+  push:
+    branches: [main]
+  schedule:
+    # run the job daily at midnight
+    - cron: "0 0 * * *"
+
+jobs:
+  bench:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v3
+
+      - name: Setup Python
+        uses: actions/setup-python@v1
+        with:
+          python-version: '3.x'
+
+      - name: Install dependencies
+        run: |
+          rustup toolchain install stable
+          rustup override set stable
+          cargo install cargo-criterion
+          pip3 install "boto3[crt]"
+
+      - name: Generate
+        working-directory: bindings/rust
+        run: ./generate.sh --skip-tests
+
+      - name: Benchmark
+        working-directory: bindings/rust/bench
+        run: cargo criterion --message-format json > criterion_output.log
+
+      - name: Configure AWS Credentials
+        uses: aws-actions/configure-aws-credentials@v1
+        with:
+          aws-access-key-id: ${{ secrets.AWS_ACCESS_KEY_ID }}
+          aws-secret-access-key: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
+          aws-region: us-west-2
+
+      - name: Emit CloudWatch metrics
+        run: |
+          python3 .github/bin/criterion_to_cloudwatch.py \
+            --criterion_output_path bindings/rust/bench/criterion_output.log \
+            --namespace s2n-tls-bench \
+            --platform ${{ runner.os }}-${{ runner.arch }}