---
title: "PostgresBench: A Reproducible Benchmark for Postgres Services"
description: "PostgresBench is a public, reproducible benchmarking suite created to compare transactional performance across managed PostgreSQL providers. The benchmark uses the standard pgbench tool with a TPC-B-like write-intensive workload executing across 256 clients and 16 threads for ten-minute durations. Evaluators tested five services—including Postgres managed by ClickHouse, AWS Aurora, AWS RDS, Neon, and Crunchy Bridge—under 100 GB and 500 GB scale factors using default configurations in the us-east-2 region. Across tested configurations, Postgres managed by ClickHouse delivered higher transactions per second and lower P99 latencies due to physically co-located NVMe primary storage. These results demonstrate that remote network storage layers, such as EBS volumes, introduce significant latency overhead during sustained write and fsync operations."
---

# PostgresBench: A Reproducible Benchmark for Postgres Services

[Clickhouse](https://yomu.fyi/company/clickhouse) · Lionel Palacin · Apr 2, 2026

**Type:** Benchmark

## Summary

PostgresBench is a public, reproducible benchmarking suite created to compare transactional performance across managed PostgreSQL providers. The benchmark uses the standard pgbench tool with a TPC-B-like write-intensive workload executing across 256 clients and 16 threads for ten-minute durations. Evaluators tested five services—including Postgres managed by ClickHouse, AWS Aurora, AWS RDS, Neon, and Crunchy Bridge—under 100 GB and 500 GB scale factors using default configurations in the us-east-2 region. Across tested configurations, Postgres managed by ClickHouse delivered higher transactions per second and lower P99 latencies due to physically co-located NVMe primary storage. These results demonstrate that remote network storage layers, such as EBS volumes, introduce significant latency overhead during sustained write and fsync operations.

## Context

Evaluating managed PostgreSQL services requires a transparent and reproducible testing methodology. Because managed providers rely on different compute and storage architectures, standard transactional comparisons are needed to assess performance without unverified vendor claims.

## Approach / What changed

PostgresBench runs standard pgbench TPC-B workloads against managed PostgreSQL services using 256 clients and 16 threads over 10-minute intervals. The benchmark evaluates two dataset sizes (~100 GB and ~500 GB) across standardized small and large instance tiers in AWS us-east-2, keeping default PostgreSQL settings and disabling high availability to isolate single-node compute and storage performance.

## Takeaways

- PostgresBench uses pgbench with 256 clients and 16 threads across scale factors 6849 (~100 GB) and 34247 (~500 GB) to measure stable throughput and disk pressure handling.
- On large instances at 100 GB scale, Postgres managed by ClickHouse reached 28,668 TPS, compared to 14,790 for Crunchy Bridge, 12,628 for AWS Aurora, 8,563 for Neon, and 8,133 for AWS RDS.
- Physically co-locating NVMe storage with compute avoids the network round-trip latency that remote storage layers like EBS incur during write-intensive WAL fsync operations.

**Tags:** [AWS](https://yomu.fyi/topic/aws), [Open Source](https://yomu.fyi/topic/open-source), [Performance](https://yomu.fyi/topic/performance), [Postgres](https://yomu.fyi/topic/postgres), [Scalability](https://yomu.fyi/topic/scalability)

- Source: [Clickhouse](https://clickhouse.com/blog/postgresbench)
- Source URL: https://clickhouse.com/blog/postgresbench
- Ingested by Yomu: 2026-08-28T01:23:41.814Z

[Read original post](https://clickhouse.com/blog/postgresbench)