My 2 cents:

> Each run lasts 10 minutes, long enough to move past warmup and capture stable throughput

But short enough not to show the effect of checkpoints (which is quite noticeable). It's not representing a realistic workload. I'd use at the very least 30mins with checkpoint_timeout=5mins so you get at least 6 checkpoint cycles.

> We report average TPS, average latency, P95 latency, and P99 latency

Unless I'm mistaken, only average TPS is reported, but not TPS over time. Again, it varies a lot (especially due to checkpoints!) and you need variation over time to understand patterns and determine SLAs.

> We tested two scale factors: 6849 (~100 GB) and 34247 (~500 GB).

I'd set them an order of magnitude apart (100GB and 1TB). 1TB+ is nowadays not unseen and therefore represents a step into the scale path that I believe should also be compared.

> While all services offer high availability, the underlying architectures differ. Some use standby replication, others use shared or distributed storage layers. Since we are focused on single-node compute and storage performance, we tested without HA enabled to isolate that.

So again, it's not realistic. The impact of HA in production is quite measurable in performance, especially if semi-sync is used.

But more importantly, you are comparing systems that are already redundant and highly available (e.g. Aurora, writing six copies of the data) with a single-instance one writing (if I'm not mistaken) to an ephemeral local NVMe. This is watermelons to grapes comparison (I assume HA CH clusters talk over the network for each confirmed transaction; but here they only go to local NVMe).

> Default Postgres configuration

I believe this is a mistake. Postgres must be tuned (at least to a reasonable degree) and try to homogenize the configuration across the tested systems. Otherwise you are not benchmarking Postgres, but rather the conservationism in the decisions of the default configurations of the managed services.

> We did not compare pricing.

So you leave the fun part out?

> Systems included

I miss a self-managed EC2 cluster (with local NVMe) for comparison.

It’s a nice project, but it doesn’t seem like the open-source community has really picked it up yet. It’s been live since May, but there aren’t many GitHub stars so far and contributors.

It would be interesting to see vanilla postgres on a VPS and vanilla postgres on a bare metal server included too. I understand it's apples and oranges, but would it be slower, would it be faster, by how much?

I'm curious to know how PlanetScale Postgres would fare.

I was just looking for something like this. I think there is a lack of benchmarks between Postgres cloud instances across different clouds. They all offer the same features mostly and cost about the same, but we don't know how they perform.

If I had ~$10,000k to burn, I would give Claude Code an AWS account and this benchmark and have it generate parameter spaces and optimize the configuration for transaction processing by just changing the configuration.

While it’s important to focus on the backend itself, why don’t anyone focus on the clients which talk to the databases, it is there most of an applications problem starts also in regards to performance.

Tldr? Can I use this to benchmark my own CNPG on k8s to see if there is improvement to be made on the deployment?