How to Calculate Apdex Score from Latency Data in Python

Generate simulated latency samples and compute the Apdex score to gauge user satisfaction with an application's performance.

Easy Python 3.9+ Aug 9, 2026 Observability & SRE 15 views 0 copies

apdex latency observability sre statistics

Python code

17 lines

Python 3.9+

import random
import statistics

def generate_latencies(count=100, base=100, stddev=30):
    return [max(0, random.gauss(base, stddev)) for _ in range(count)]

def apdex(latencies, threshold=200):
    satisfied = sum(1 for lat in latencies if lat < threshold)
    tolerating = sum(1 for lat in latencies if lat >= threshold and lat < threshold * 4)
    return (satisfied + tolerating / 2) / len(latencies) if latencies else 0

if __name__ == "__main__":
    random.seed(42)
    lat = generate_latencies()
    print(f"Latencies: {lat[:5]}...")
    print(f"Apdex score: {apdex(lat):.3f}")
    print(f"Avg latency: {statistics.mean(lat):.1f}ms")

Output

stdout

Latencies: [426.8685625423253, 76.15577676701064, 107.99627901091701, 165.59748845021215, 113.35493176170674]...
Apdex score: 0.975
Avg latency: 99.7ms

How it works

The Apdex formula maps each latency to one of three buckets: satisfied (< threshold), tolerating (>= threshold but < 4x threshold), and frustrated (all others). The final score is (satisfied + tolerating/2) / total, giving 1.0 when all requests meet the goal. Using a seeded random generator makes the output reproducible, which is useful for unit tests. The max(0, ...) guard avoids negative latency samples from the Gaussian distribution.

Common mistakes

Forgetting to seed the random generator, causing non-reproducible results
Using >= vs > inconsistently in bucket boundaries, skewing the score
Dividing by zero when the latency list is empty
Using threshold * 4 correctly — it must be 4x, not 2x, the threshold

Variations

Use numpy to vectorize the bucket counting for large datasets
Compute Apdex from real latency logs instead of simulated data

Real-world use cases

Monitoring API response times in a microservices dashboard to alert on poor user experience.
Evaluating the impact of a new caching layer by comparing Apdex scores pre- and post-deployment.
Setting SLOs (service level objectives) for latency-based performance targets in production.

How to Calculate Apdex Score from Latency Data in Python

Python code

Output

How it works

Common mistakes

Variations

Real-world use cases

More from Observability & SRE

Tutorials

Quizzes

Python code

Output

How it works

Common mistakes

Variations

Real-world use cases

More from Observability & SRE

Keep learning

Tutorials

Quizzes