(benchmark): Add a script to analyse the log data

Author: alex9smith

.gitignore

@@ -1,3 +1,4 @@
 benchmark/data/*
+benchmark/scripts/plots/*
 __pycache__/*
 .DS_Store

TODO.md

@@ -3,16 +3,10 @@
 ## Lambdas
 
 - Python
-- Ruby
 - Java
 
 - Use futures / async to write to DynamoDB concurrently in the Rust lambda
 
-## Infrastucture
-
-- Deploy the other lambdas
-
 ## Benchmark
 
-- Initial analysis - see what's there
-- Script for reproducible analyses
+- Redo analysis with all the Lambda data

benchmark/scripts/analyse_results.py

@@ -0,0 +1,71 @@
+import seaborn as sns
+import matplotlib.pyplot as plt
+import pandas as pd
+import numpy as np
+
+sns.set_theme(style="whitegrid", palette="colorblind")
+
+
+def save_plot(ax, plot_name: str) -> None:
+    figure = ax.get_figure()
+    figure.savefig(f"plots/{plot_name}.png")
+    plt.close(figure)
+
+
+def print_and_plot_init_duration_by_language(data: pd.DataFrame) -> None:
+    cold_starts = data[data["cold_start"] == True]
+
+    print("Cold start times per language - p50, p95, p99")
+    print(
+        cold_starts.groupby("name")["init_duration_ms"]
+        .quantile(q=np.array([0.50, 0.95, 0.99]))
+        .unstack()
+    )
+    print("-" * 10)
+
+    ax = sns.histplot(x="init_duration_ms", hue="name", binwidth=20, data=data)
+    ax.set_title("Cold start times by language")
+    ax.set_xlabel("Init duration (ms)")
+    ax.set_ylabel("Count")
+    ax.set_xlim(0)
+    sns.despine()
+    save_plot(ax, "init_duration_by_language")
+
+
+def print_and_plot_run_time_by_language(data: pd.DataFrame) -> None:
+    print("Execution times per language - p50, p95, p99")
+    print(
+        data.groupby(["name", "cold_start"])["execution_time_ms"]
+        .quantile(q=np.array([0.50, 0.95, 0.99]))
+        .unstack()
+    )
+    print("-" * 10)
+
+    ax = sns.histplot(x="execution_time_ms", hue="name", binwidth=5, data=data)
+    ax.set_title("Execution times by language")
+    ax.set_xlabel("Execution time (ms)")
+    ax.set_ylabel("Count")
+    ax.set_xlim(0, 500)
+    sns.despine()
+    save_plot(ax, "run_time_by_language")
+
+
+def print_total_cost_by_language(data: pd.DataFrame) -> None:
+    # From August 1st 2025 AWS will charge for the INIT phase
+    # https://aws.amazon.com/blogs/compute/aws-lambda-standardizes-billing-for-init-phase/
+    data["total_time"] = data["execution_time_ms"] + data["init_duration_ms"]
+    totals = data.groupby("name")[
+        ["init_duration_ms", "execution_time_ms", "total_time"]
+    ].sum()
+    totals["init_percentage"] = totals["init_duration_ms"] / totals["total_time"]
+
+    print("Total billed time by language")
+    print(totals)
+    print("-" * 10)
+
+
+if __name__ == "__main__":
+    data = pd.read_csv("../data/parsed_cloudwatch_logs.csv")
+    print_and_plot_init_duration_by_language(data=data)
+    print_and_plot_run_time_by_language(data=data)
+    print_total_cost_by_language(data=data)

benchmark/scripts/requirements.txt

@@ -1,7 +1,20 @@
 boto3==1.38.6
 botocore==1.38.6
+contourpy==1.3.2
+cycler==0.12.1
+fonttools==4.58.0
 jmespath==1.0.1
+kiwisolver==1.4.8
+matplotlib==3.10.3
+numpy==2.2.5
+packaging==25.0
+pandas==2.2.3
+pillow==11.2.1
+pyparsing==3.2.3
 python-dateutil==2.9.0.post0
+pytz==2025.2
 s3transfer==0.12.0
+seaborn==0.13.2
 six==1.17.0
+tzdata==2025.2
 urllib3==2.4.0